Vol 20 Num 2

MID-AMERICAN JOURNAL OF BUSINESS
Fall 2005
Volume 20, Number 2
3 7
EDITORIAL
Managing Our Way to Academic Decline
Ashok Gupta
DEANS FORUM
Business Colleges Should Practice What They Preach
Daniel Vetter
EXECUTIVE VIEWPOINT
Corporate Culture Denes a Company and Its Future
Richard T. Farmer
11 21 25 35
ARTICLES
An Analysis of Working Capital Management Results Across Industries
Greg Filbeck and Thomas M. Krueger
IBBEA Implementation and the Relative Protability of Small Banks

Srinivas Nippani and Kenneth M. Washer
Modeling Internet Operations Using Initial Public Offerings

Sameer Prasad, David C. Porter, and Linda Yu
The Communication Effectiveness of System Models Using the UML versus Structured Techniques: A Field Experiment
Bruce C. Hungerford and Michael A. Eierman
45 59 70
Building a Market-Oriented Organizational Environment: An Implementation Framework for Small Organizations

Beth Ann Martin and James H. Martin
Implementing Planned Change: An Empirical Comparison of Theoretical Perspectives

Matthew W. Ford and Bertie M. Greer
BOOK REVIEW
Final Accounting
Bill Cummings
www.bsu.edu/majb
Gupta
EDITORIAL
Managing Our Way to Academic Decline

One of the critical elements with a signicant impact on the quality of education is the recruitment and retention of excellent faculty. In a recent article, Bennis and OToole lamented that business schools lost their way by focusing on the wrong reward system for faculty, which emphasizes rigorous-scientic rather than practical-relevant research. Top rated business schools derive pleasure and prestige by extolling the number of articles their faculties have published in A-rated journals. There is nothing wrong in publishing in A-rated journals; but there is something seriously wrong when the consequences of such behavior on business education are ignored; it is even more deplorable when not-so-top-rated business schools try to emulate their A-rated brethrens. We have long known the perils of rewarding behavior A while hoping for behavior B; we must now deal with the outcome of such thinking. Ineffective management of business schools is a major source of discontent among faculty. It is paradoxical and ironic when business schools that are supposed to teach how to manage a business are themselves poorly managed. Deans come and go as in a revolving door. Jeffrey Garten, departing Dean of the Yale School of Management, said in a recent New York Times interview, I was an investment banker for fteen years. I was in four presidential administrations. But this job has been the most difcult of all. Looking at the number of ads for Deans jobs, one wonders where all the good people have gone! Many business schools have hired (tired or retired) corporate executives as Accidental Deans deans positions occupied by those with no training or experience in teaching, research or academic leadership. They do not know and perhaps dont care what it means to be a faculty in a university setting. Many have no appreciation or respect for faculty governance in higher education. They are used to a corporate style of management and that is how they run the business school and treat business faculty. Ethics for them is something to organize lectures on and write about. In the name of discretionary powers, they play favorites with faculty without regard to the demoralizing effect such actions may have on others. In the name of raising money, they enjoy traveling and meeting their own corporate types with whom they are most comfortable. Extensive traveling as a by-product, keeps them away from campus which helps them avoid interacting with faculty and getting to know them except, of course, their own core group of condants called the members of the Executive Council, which keeps getting bigger and bigger with additional Assistant and Associate Deans. In the name of innovation, they dilute the curriculum without assessing the effectiveness of such an innovation. As if grade ination was not enough, some business schools have created lucrative programs for students to receive multiple credits for little work. These innovations essentially give luster to the students resume a truly studentcentered approach to education! Why are we just concerned about rigor in research and not rigor in education? Faculty members, on the other hand, have become timid. Like Pavlovs dog, they have learned how to behave: give easy grades faculty gets good evaluations, students feel good (they are already paying steep tuition why give them a hard time) and faculty do not have to defend their actions; just say yes and grease the right administrators to get lucrative assignments; volunteer to get on the important recruitment committees; become a showman invite the Dean or other big-wigs during student presentations where you exhibit well dressed executive-looking students and their slick PowerPoint slides rather than focus on the content of their presentations or grill them with tough questions; and if you are a senior professor, you 3
Ashok Gupta Editor-in-Chief
Mid-American Journal of Business, Vol. 20, No. 2
Gupta
collaborate with some untenured faculty members or, if you are lucky, with young relatives on research projects and publications preferably for A-journals! Students are happy with less. Education is perhaps the only industry where the customer is least demanding and happy with little. They prefer a lighter work-load, easy grading, entertaining classes, curving of grades, and accommodating professors. Many students work to pay for their education; they have to t classes into their real lives. A large number come from difcult family backgrounds with poor study habits and work ethics. They consider getting a diploma as their entitlement for paying the tuition. Students attitudes toward learning need serious adjustments. In a truly global world that we now live in, American students are not just competing with other fellow students in America; they are competing with students from rest of the world. To stay competitive, US students need to keep getting better. Thomas Friedman of the New York Times beautifully summarized how things have turned around: When I was growing up, my parents used to say to me, Tom, nish your dinner people in China are starving. But after sailing to the edges of the at world for a year, I am now telling my own daughters, Girls, nish your homework people in China and India are starving for your jobs. [3] There was a time when the world used to knock at Americas doors for excellence in higher education in science, mathematics, engineering, technology, and business. America still attracts the best talent from around the world. However, that lure may be slowly diminishing. Students around the world now have more choices; those who come to America for higher education may prefer to return to their native lands, taking their talent with them. India presents a shining example of academic excellence in several of these areas of education. India is setting up campuses around the world to provide rigorous and relevant education at a lower price-tag. Will we be ready to compete? Lets rededicate ourselves to the real purpose of business education: producing students who can generate creative solutions to business problems in the absence of clear facts by integrating knowledge, experience and critical thinking. Lets remember that rigor and relevancy need not be mutually exclusive. We can create a rigorous and relevant business curriculum, and at the same time conduct practical-relevant business research in a scrupulous and scientic manner. These goals can only be achieved with the help of academic leaders who truly understand the role of faculty in business education and who treat them with respect, fairness and dignity.
CFO magazines annual Working Capital Management Survey, attempt to answer two questions: (1) are rms in one industry as opposed to another quickly able to transfer sales into cash; and (2) does working capital management performance for rms within a given industry change from year-to year? The authors report an afrmative answer to both questions. The second article examines the impact of the Interstate Banking and Branching Efciency Act (IBBEA) of 1994 on the performance of small banks relative to large banks. The authors compare the performance of small and large banks in the periods preceding and following (1988-2002) IBBEA implementation. The study concludes that IBBEA has put small banks at a competitive disadvantage. The sudden Dot Com boom and bust left many investors and venture capitalists wondering about the investment worthiness of Internet companies. In the third article of this issue, the authors examine 340 Initial Public Offerings (IPO) of Internet companies to identify which types of companies are likely to have superior performance. The authors nd that Internet rms with high information intensity and low customer contact yield superior performance but rms with low physical presence underperformed. The study reported in the fourth article examines the comparative effectiveness of the Unied Modeling Language (UML) and traditional modeling languages in communicating information about a system design. Effectiveness is assessed by examining the performance of three groups: (1) individuals with no knowledge of either language; (2) individuals with no knowledge of either modeling language that were provided training in one of the languages; and (3) individuals that have had more extensive training in one of the languages. Although the impact of market-orientation on organization performance has now been widely accepted, there is little guidance for actually developing market-orientation. In the fth article of this issue, the authors, using an internal customer-internal supplier perspective, identify a framework for creating a market-oriented workforce in small organizations where impact of market-orientation on performance is stronger than in the larger organizations. Few models for implementing planned change have been studied using empirical research designs. In the last article, using data from over one hundred managers involved in the implementation of planned change, the authors draw conclusions about the appropriateness of three congurations and about the relative importance of various change process factors in achieving implementation success.
In this issue......
In this issue, we present a set of six articles and an executive viewpoint. Businesses are paying increased attention to applying Six Sigma methodologies to measure and ensure quality in all areas of the enterprise including working capital management. In the rst article of this issue, the authors, using 4
Mid-American Journal of Business, Vol. 20 No. 2
References
Bennis, Warren G. and James OToole, How Business Schools Lost Their Way, Harvard Business Review, May, 2005. Holstein, William J., Are Business Schools Failing the World? The New York Times, June 19, 2005. Friedman, Thomas, Its a Flat World, After All, New Yourk Times, April 3, 2005
Vetter
DEANS FORUM
Business Colleges Should Practice What They Preach

As I was facilitating discussion regarding the colleges three year strategic plan at our recent Deans Business Advisory Board meeting in May, a board member inquired whether the college practices what it preaches. We teach our students how successful rms and organizations operate, but do we expect that of ourselves? He also suggested, If we dont, we ought to. He introduces a compelling point. Is a business college a good example of how a business is run? The answer is yes and no. Daniel Vetter A business college is much like a busiInterim Dean ness with products and services, clients College of Business or customers, and business processes. Administration Central Michigan University Granted, to the extent that business colleges can act like a business in all aspects is somewhat limited. However, there are some best business practices that we can follow and possibly improve on. chairpersons and program directors are key, but faculty are the biggest challenge. Not only does a dean have to nd a way to include the faculty into the strategic planning process, but faculty activities such as teaching, research, professional development, and service must align with college priorities. Faculty must have incentives to focus their activities on college priorities. They must understand the importance of what they do and how it contributes to the college mission and vision.
Become more entrepreneurial

Some business colleges throughout the country have done this well. For the most part, we can all improve. Larger and more prominent business colleges have a strong track record of courting alumni, pursuing private money, aggressively raising fees to cover program costs, and growing executive education programs. We all can learn from this type of model. Our college still has considerable progress to make in this area. However, we have been more aggressively leveraging business partnerships to help sponsor programs and focus initiatives. Our faculty do not have expertise in every area, but we are getting better at marketing and matching our focused expertise to offer management education programs that rms and organizations value.
Live your strategic plan

Create and communicate your vision for the college. Take the college with you. AACSB Internationals emphasis on strategic planning has been very good for business colleges. For some business colleges, the strategic planning process is standard fare. For others it is new territory. Gradually bringing faculty and staff together to build a consensus about where the college is going, providing benchmarks from which progress can be measured, and focusing nancial resources on key initiatives are key components. Our college has formally employed a strategic planning process for a number of years. We always struggle with several issues. Consistency with ever changing university goals is one challenge. Bringing a college organization together to embrace the importance of the strategic plan is the greatest challenge. My experience suggests that the college staff are the easiest to embrace the process. Department
Measure how well your college is performing

Benchmarking is a time honored practice. Comparing our performance to our peers has been healthy. We all have a few schools that we aspire to be like. Nevertheless, business schools should also look to campuses that are different and who may have useful and smart innovative programs. Although university presidents are enamored with rankings, in my opinion, we spend too much time and resources to
Vetter
pursue rankings. Some of the ranking criteria is sensible, but perceptions are not a good way to rank schools. We need to be more objective and relevant. More across the board benchmarking is needed in the area of student learning. What is the real value that a business education provides? The proof is in the pudding. It is in the performance of our graduates. Do they have the skills, knowledge, and abilities to compete? We also have to improve on communicating these results to our stakeholders.
The Mid-American Journal of Business

is sponsored by:
Ball State University
Miller College of Business Lynne Richardson, Dean
Improve the skills of your workforce

Are we doing enough to incentive and motivate faculty to continuously maintain and upgrade professional development skills? Are scarce resources spent effectively? We usually rely on faculty to personally make these decisions. Do they make the correct decision? In the area of scholarship, some faculty research skills are steadily eroded after graduate school. Should this happen? I believe it is somewhat universal that senior faculty research agendas change. However, I would like to challenge our faculty to continue to do work that makes a signicant impact on their discipline and is consistent with the college mission. What are the incentives and programs required to maintain the rigor and level of faculty scholarship that they once produced?
Central Michigan University

College of Business Administration Daniel Vetter, Interim Dean
Miami University
Richard T. Farmer School of Business Roger Jenkins, Dean
Know your customer and markets

Deans spend considerable time connecting with college stakeholders. An article in the May/June 2004 BizEd magazine highlights the ability of small business schools to know their markets and nd niches that they can exploit. Perfecting a single program that is attractive to your state or region is much more manageable than a number of different programs. It is also more desirable to retain your current customers rather than have to attract new ones. To be aware of what is relevant and timely, faculty need to connect more with the business community on a regular basis. This is especially important for new program development. Business college programs and new curriculum development must be more responsive. We still cannot respond quickly enough to a changing business environment. The great business colleges do it. However, business colleges tend to take years to revise business degree programs. Miami University totally redesigned their MBA program in a very short period of time. Thats a good example of the results that we must see from business colleges. We must be aware that we are now in a world of continuous development of degree programs. The days of the stagnant, plain vanilla degree program are long gone. In conclusion, I ask deans, associate deans, directors, department chairs, and faculty; if you were asked whether you practice what you preach by an advisory board member, employer, or business professional, how would you answer the question? Could you do better? 6
Northern Illinois University

College of Business William Tallon, Interium Dean
Ohio University
College of Business Glenn Corlett, Dean
The University of Toledo

College of Businss Administration Thomas Gutteridge, Dean
Western Michigan University

Haworth College of Business Adrian Ed Edwards, Interim Dean
Farmer
EXECUTIVE VIEWPOINT
Corporate Culture Denes a Company and its Future

Cintas Corporation is known in business and investment circles as a performer. Our record of thirty-ve consecutive years of growth in sales and prots is nearly unheard of; according to our research, only Wal-Mart can match it. We routinely are included in the business elite, being heralded by Fortune magazine as one of Americas Most Admired Companies, by Forbes as one of Americas Best Managed Companies and by Mergent as a Dividend Achiever. While we are extremely proud of these achievements, we are even more proud of what drives them: The Cintas culture. Corporate culture is what separates the business winners from the business losers. A corporate culture of honesty and integrity is more valuable than cash in the bank. It is the glue that holds a company together and helps companies like Cintas do great things. Quite simply, our culture is our No. 1 competitive advantage. At Cintas, our culture is no accident. It is our planned approach to our business. We spend time and money teaching our culture, protecting it and perpetuating it. We are condent our resources are well spent in that regard. Safeguarding our culture is safeguarding our future. are having a positive effect or not, the plain fact is that Sarbanes-Oxley will not solve the problem of corporate greed. The real problem is a lack of honesty and integrity on the part of some people running some companies. You cant legislate honesty and integrity. You can, however, make honesty and integrity part of your corporate culture. You can hire those traits, you can breed them, you can reward them. You can make high standards as much a part of your business as the bricks on the building or the name on the door.
Richard T. Farmer Chairman and Founder, Cintas Corporation
A Culture of Opportunity
At Cintas, our culture starts with our principal objective, which is to maximize the long-term value of Cintas for its shareholders and working partners by exceeding our customers expectations. We base every decision on that objective. In one sentence, it summarizes our ongoing purpose for being. In following our principal objective at every level of our company, we can be sure that decisions are made in the best interest of our company and our working partners, that is, our employees. We dont take short-term gains at the expense of long-term values, ever. Organizations that do not have a guiding principle or that do not follow it, often end up being run for the benet of top management. Examples of such companies abound. They borrow money they cant pay back. They take the easy way out on union contracts and they make bad acquisitions. They go for ego-feeding notoriety rather than long-term stability. Some dont explore new opportunities because theyre pretty darned comfortable right where they are. In time, their bad decisions catch up with them and the companies fail.
Why Culture Counts

The headlines over the past few years have called out a myriad of corporate misdeeds, and rightly so. Unethical behavior has cost investors billions of dollars and created an atmosphere of suspicion and mistrust on the part of employees, shareholders and the public at large. The government responded with a series of regulations that themselves carry a hefty price tag for companies and therefore for investors. Without judging whether the regulations are right or wrong, or whether they
Farmer
The companies that are accused of overstating sales or prots, of paying outrageous salaries, bonuses and stock options, clearly were not working to maximize the longterm value of their businesses for their shareholders or working partners. They were looking out for the interests of a few. With everyone at Cintas working to maximize the longterm value of the company for shareholders and working partners, we are creating career and growth opportunities. With everyone at Cintas working to exceed our customers expectations, we are building our reputation and our business. Our visionand we communicate it daily throughout our companyis to provide a product or service to every business in North America. Under the guidance of the principal objective, keeping our customers at the center of our universe, we believe we can attain that vision.
Respect also is evident in the fact that we live by the rules and we recognize that no one is above them. We strictly separate business and personal affairs, and refrain from any personal activity that could have the appearance of inuencing business decision-making. We have a direct line to top management for partners to share any ethical concerns. We strive to keep our relationships and our environment professional at all times. In fact, if there is one word that embodies the character of Cintas partners, it is professional...in every sense of the word. We are professional in the way we dress, the way we act, the way we treat one another and the way we conduct our business. Our corporate tagline reects that, too. It is The Service Professionals.
A Culture of Leadership
In addition to living a culture of opportunity and respect, we believe in a strong culture of leadership. One of our executives once said, Good leaders start with the answer. By this he meant that good leaders decide where the organization needs to be, what it wants to accomplish and then do what it takes to get there. Our principal objective, by its very nature, encourages leadership that is both strong and ethical. A rm set of rules helps, too. Back in the early days of Cintas, I was the only manager. I had two routes and twelve employees. I did it all...checking on the drivers, selling accounts, collecting them, handling service issues and managing our growth. And we did grow. In time, I promoted people to supervisory jobs, helping them along the way, telling them what to do and how to do it. But I didnt put it in writing, not at rst. I told partners how to handle the ordinary, but not how to handle the extraordinary. Partners didnt know what to do with unusual requests or if problems interfered with their normal way of doing their jobs. Partners began to improvise. They didnt understand how changes in their departments affected partners in other areas of the company. Standards changed. Systems changed. And so I went back to the drawing board, and I mean the drawing board. I put everything in writing: A system of policies and guidelines along with a system for creating or revising policies. Then I meticulously enforced those guidelines. In that way, I put to work years of experience, a common dialogue and, really, a leadership perpetuation plan. Together, we created a self-regulating organization in which everyone knows what he or she can and cannot do. We created a culture of accountable leadership, strengthened by a communicated vision and a clear set of parameters. We gave our leaders the tools they need to do their jobs. Cintas leaders walk the talk. They live the culture. They are dependable and trustworthy, competent and committed. They expect the same of their partners. Our leaders also are consistent in their decisions and in their training.
A Culture of Respect
In our business, the customer is king. We dont just want to satisfy customers; we want to make them big fans of Cintas. We are all on that page, everyone at our company. The only kinds of partners we have are those who demonstrate the importance of exceeding our customers expectations every day. Thats important to know. It says a lot about who we are and how we grew. The roots of respect were planted even before there was a Cintas. When I was a boy, I used to go with my mom and dad to my grandpas company, Acme Industrial Laundry. Id fall asleep in a basket of warm, clean and cozy towels while my parents worked shoulder to shoulder with employees. That was the beginning of the Cintas culture, even though no one had ever heard of the term back then. The fact was that by working side by side with employees, by eating, laughing and talking together, we learned about their lives. We respected and appreciated them as individuals and as valuable partners with valuable ideas on how to do things better. As the company grew, so did this spirit. It stayed with us as Acme evolved and then became Satellite and then became Cintas. We learned by doing, we stayed close to our partners, and I can honestly say we never got too big for our britches. Each of the three CEOs who have led our company spent time early in his career driving trucks and working the plant oors. We work as a team, respecting partners and their roles in the success of the company. If we are buying new trucks, the partners driving them are involved in the decision. If we are considering buying a piece of machinery, the operators help decide what well do. Who better to give advice than the people who do the work? That kind of respect has helped our company avoid costly mistakes while creating a stronger sense of ownership among all of our partners. 8
Farmer
They push, they nourish, they inspire. They create and maintain a team spirit that is larger than any individual, larger than any business unit.
A Culture of Excellence
At Cintas, the pace is quicker; the intensity higher; the expectations greater. We work hard with a sense of competitive urgency, a thoroughness and enthusiastic attention to detail. We also share a sense of positive discontent, rmly believing that no matter how good things are, they can always be better. We constantly strive to improve our processes, our systems, our products and our services. Exceeding customers expectations and driving longterm value is the simple, overriding business necessity. That is the attitude required to compete, and we pursue it with a passion you can feel. Think about a parent or a coach or a boss who pushed you or stretched what you thought were your limits. In expecting more, they helped you accomplish more. They wouldnt accept anything but the best, and so you delivered. A company can and should create that same culture of excellence.
Farmers early vision laid the groundwork for the company that Cintas is today. Over the years, he has served the company in many capacities, including president and CEO. He has been honored as Ernst & Youngs Entrepreneur of the Year and twice was named CEO of the Year by Financial World magazine. In 2005, Cintas was named to Fortune magazines list of Americas Most Admired Companies for the fth consecutive year. Cintas has grown in sales and prots for 35 consecutive years, through all economic cycles. With sales approximating $3 billion in FY 05 and more than 30,000 employee-partners at over 350 locations nationwide, Cintas is a publicly held company traded over the NASDAQ National Market under the symbol CTAS.
A Culture of Success
And so by using honesty and integrity as building blocks, and having a clear vision clearly communicated, Cintas has created a culture of opportunity, respect, leadership and excellence. We also have created a certain spirit that permeates our entire organization, a spirit that has driven a culture of success for our partners and our company. The spirit is the difference. If, as a business or thought leader, you are in a position to strengthen your corporate culture, do it. Build everything around absolute honesty and integrity. Work hard to dene your values, to put them into writing, to communicate them, to teach them, to live them. Make sure every person at every level of your organization knows the standardand the consequences of not following it. As a leader, there is no greater task you can perform to better ensure that your company is a winner not just for today, but also over the long term. About the Author
Richard T. Farmer is founder and chairman of the board of Cintas Corporation. Headquartered in Cincinnati, OH, Cintas provides highly specialized services to businesses of all types throughout North America. Cintas designs, manufactures and implements corporate identity uniform programs and provides entrance mats, restroom supplies, promotional products, rst aid and safety products, re protection services and document management services for approximately 700,000 businesses.
Farmer
MIDWEST DECISION SCIENCES INSTITUTE
2006 Annual Meeting

Crowne Plaza HotelIndianapolis, IN
March 23-25, 2006
Call for Papers

The program committee invites you to submit: Refereed Competitive Papers: Submissions are accepted in this category if the author certies that the paper has not been copyrighted, published, or presented (or submitted for presentation) at another professional meeting. Anyone submitting a paper must intend to register for and attend the meeting if the paper is accepted. Abstracts: Abstracts will be reviewed by the track chairs and will be scheduled for presenters as allowed by the meeting schedule. Must register and attend the meeting. Student Papers: Papers must be solely of student authorship. The meeting registration fee will be waived for students whose papers are accepted. Student papers will be subject to the review process and if accepted will be scheduled on the program. Must attend to be eligible for awards given at luncheon on March 24. Symposia, Tutorials, and Workshops: Submit a 2+ page summary of your proposed session and why it is of interest and importance to institute members. Follow the guidelines below for submissions. While considerable latitude and exibility as to content and conduct is allowed, you must provide your own participants (such as panel members) and any unusual equipment needs. Must register and attend the meeting.
Volunteers
Paper reviewers, discussants, and session chairs will be needed. Please contact Carl Briggs (briggs@indiana.edu
CALL FOR PAPERS
Membership Requirement
Awards will be given for Best Paper and Best Student Paper at the meeting. Student Papers must be authored by students only and identied as eligible to compete.
Instructions for Contributors

All submissions are due by January 17, 2006. Notication of acceptance or rejection will be emailed the week of February 20, 2006. Event information, registration and submissions can be made at www. mwdsi2006.org
Tracks: Accounting and Finance Global Business Management and Strategy Information Technology Innovative Education Operations Manufacturing and Services Prefessional Practice Quantitative Methods and Statistics, and Simulation Modeling Supply Chain and Marketing Management Student Papers
BEST PAPER AWARD

Receive $500 cash award and opportunity to be published in Mid-American Journal of Business.
Submission Deadline: January 17, 2006
10
An Analysis of Working Capital Management Results Across Industries

Greg Filbeck, Schweser Study Program Thomas M. Krueger, University of Wisconsin-La Crosse
Abstract
Firms are able to reduce nancing costs and/or increase the funds available for expansion by minimizing the amount of funds tied up in current assets. We provide insights into the performance of surveyed rms across key components of working capital management by using the CFO magazines annual Working Capital Management Survey. We discover that signicant differences exist between industries in working capital measures across time. In addition, we discover that these measures for working capital change signicantly within industries across time.
sales are outstanding, resulting in an increased cash ow of approximately $2 million at Thibodaux Regional Medical Center. Furthermore, bad debts declined from $3.4 million to $600,000. However, Waxers (2003) study of multiple rms employing Six Sigma nds that it is really a get rich slow technique with a rate of return hovering in the 1.2 4.5 percent range.
Even in a business using Six Sigma methodology, an optimal level of working capital management needs to be identied.
Introduction
The importance of efcient working capital management (WCM) is indisputable. Working capital is the difference between resources in cash or readily convertible into cash (Current Assets) and organizational commitments for which cash will soon be required (Current Liabilities). The objective of working capital management is to maintain the optimum balance of each of the working capital components. Business viability relies on the ability to effectively manage receivables, inventory, and payables. Firms are able to reduce nancing costs and/or increase the funds available for expansion by minimizing the amount of funds tied up in current assets. Much managerial effort is expended in bringing non-optimal levels of current assets and liabilities back toward optimal levels. An optimal level would be one in which a balance is achieved between risk and efciency. A recent example of business attempting to maximize working capital management is the recurrent attention being given to the application of Six Sigma methodology. Six Sigma methodologies help companies measure and ensure quality in all areas of the enterprise. When used to identify and rectify discrepancies, inefciencies and erroneous transactions in the nancial supply chain, Six Sigma reduces Days Sales Outstanding (DSO), accelerates the payment cycle, improves customer satisfaction and reduces the necessary amount and cost of working capital needs. There appear to be many success stories, including Jennifer Townes (2002) report of a 15 percent decrease in days that
Even in a business using Six Sigma methodology, an optimal level of working capital management needs to be identied. Industry factors may impact rm credit policy, inventory management, and bill-paying activities. Some rms may be better suited to minimize receivables and inventory, while others maximize payables. Another aspect of optimal is the extent to which poor nancial results can be tied to sub-optimal performance. Fortunately, these issues are testable with data published by CFO magazine (Mintz and Lazere 1997; Corman 1998; Mintz 1999; Myers 2000; Fink 2001), which claims to be the source of tools and information for the nancial executive, and are the subject of this research. In addition to providing mean and variance values for the working capital measures and the overall metric, two issues will be addressed in this research. One research question is, are rms within a particular industry clustered together at consistent levels of working capital measures? For instance, are rms in one industry able to quickly transfer sales into cash (i.e., have low accounts receivable levels), while rms from another industry tend to have high sales levels for the particular level of inventory (i.e., a high inventory turnover). The other research question is, does working capital management performance for rms within a given industry change from year-to-year? The following section presents a brief literature review. Next, the research method is described, including some information about the annual Working Capital Management Survey published by CFO magazine. Findings are then presented and conclusions are drawn.
11
Filbeck and Krueger
Table 1 Industries Represented in CFOs Working Capital Management Surveys
Aerospace Apparel Beverages Building Materials Chemicals Conglomerates Electric & Gas Utility Electrical Equipment Food Food & Drug Stores Food Services Forest & Paper Products Furniture General Merchandisers Health Care Health-Care Equipment
Household Products Metal Products Metals Motor Vehicles & Parts Ofce Equipment Petroleum Pharmaceuticals Publishing & Printing Recreational Scientic Equipment Semiconductors Specialty Retailers Telecommunications Textiles Transportation Wholesale Trade
Related Literature
The importance of working capital management is not new to the nance literature. Over twenty years ago, Largay and Stickney (1980) reported that the then-recent bankruptcy of W.T. Grant, a nationwide chain of department stores, should have been anticipated because the corporation had been running a decit cash ow from operations for eight of the last ten years of its corporate life. As part of a study of the Fortune 500s nancial management
practices, Gilbert and Reichert (1995) nd that accounts receivable management models are used in 59 percent of these rms to improve working capital projects, while inventory management models were used in 60 percent of the companies. More recently, Farragher, Kleiman and Sahu (1999) nd that 55 percent of rms in the S&P Industrial index complete some form of a cash ow assessment, but did not present insights regarding accounts receivable and inventory management, or the variations of any current asset accounts or liability accounts across industries. Thus, mixed evidence exists concerning the use of working capital management techniques. Theoretical determination of optimal trade credit limits are the subject of many articles over the years (e.g., Schwartz 1974; Scherr 1996), with scant attention paid to actual accounts receivable management. Across a limited sample, Weinraub and Visscher (1998) observe a tendency of rms with low levels of current ratios to also have low levels of current liabilities. Simultaneously investigating accounts receivable and payable issues, Hill, Sartoris, and Ferguson (1984) nd differences in the way payment dates are dened. Payees dene the date of payment as the date payment is received, while payors view payment as the postmark date. Additional WCM insight across rms, industries, and time can add to this body of research. Maness and Zietlow (2002, 51, 496) presents two models of value creation that incorporate effective shortterm nancial management activities. However, these models are generic models and do not consider unique rm or industry inuences. Maness and Zietlow discuss industry
Table 2 Working Capital Management Component Denitions and Averages

Component Equation Average Value (Standard Deviation) 9.0 percent (1.7 percent) 51.8 days (4.7 days)
Cash Conversion Efciency (CCE)
(Cash ow from operations) / Sales
Days Working Capital (DWC)
(Receivables + Inventory Payables) / (Sales/365) (Highest overall CCE Company CCE) / (Highest overall CCE Lowest overall CCE) x (Lowest overall DWC Company DWC) / Lowest overall DWC Highest overall DWC)
Overall Ranking
Other Related Variables Those listed below, although reported in CFO, are not part of the overall ranking criteria (only the two meaures listed above are included in overall rank). Days Sales Outstanding Accounts Receivable / (Sales/365) 50.6 days (1.3 days) 11.0X/year or 32.4 days (2.5 days) 32.0 days (2.8 days)
Inventory Turns
Inventory / (Sales/365)
Days Payables Outstanding
Accounts Payable / (Sales/365)
12
Filbeck and Krueger
inuences in a short paragraph that includes the observation that, An industry a company is located in may have more inuence on that companys fortunes than overall GNP (2002, 507). In fact, a careful review of this 627-page textbook nds only sporadic information on actual rm levels of WCM dimensions, virtually nothing on industry factors except for some boxed items with titles such as, Should a Retailer Offer an In-House Credit Card (128) and nothing on WCM stability over time. This research will attempt to ll this void by investigating patterns related to working capital measures within industries and illustrate differences between industries across time. An extensive survey of library and Internet resources provided very few recent reports about working capital management. The most relevant set of articles was Weisel and Bradleys (2003) article on cash ow management and one of inventory control as a result of effective supply chain management by Hadley (2004).
measures of working capital efciency identied by CFO magazine. Classical analysis of variance is used to address issues of industry rank differences within years. Thus, H 1: Differences exist among industries with respect to the measures of working capital efciency identied by CFO magazine. Our second hypothesis is that working capital measures for rms within an industry change across time. Since the complete data set includes only four years (19961999), there is the potential for degrees of freedom issues when using sophisticated models. Assessment of WCM performance across years is conducted using the Kendalls Coefcient of Concordance. Thus, H 2: Working capital measures for rms within an industry change across time.
Research Method
The CFO Rankings The rst annual CFO Working Capital Survey, a joint project with REL Consultancy Group, was published in the June 1997 issue of CFO (Mintz and Lezere 1997). REL is a London, England-based management consulting rm specializing in working capital issues for its global list of clients. The original survey reports several working capital benchmarks for public companies using data for 1996. Each company is ranked against its peers and also against the entire eld of 1,000 companies. REL continues to update the original information on an annual basis. The industries that include at least eight companies with complete information over the 1996-2000 period are listed in Table 1. REL uses the cash ow from operations value located on rm cash ow statements to estimate cash conversion efciency (CCE). This value indicates how well a company transforms revenues into cash ow. A days of working capital (DWC) value is based on the dollar amount in each of the aggregate, equally-weighted receivables, inventory, and payables accounts. The days of working capital (DNC) represents the time period between purchase of inventory on acccount from vendor until the sale to the customer, the collection of the receivables, and payment receipt. Thus, it reects the companys ability to nance its core operations with vendor credit. A detailed investigation of WCM is possible because CFO also provides rm and industry values for days sales outstanding (A/R), inventory turnover, and days payables outstanding (A/P). More information on how these values are calculated is presented in Table 2. Prior to 2002, CFO also provided an overall WCM management ranking based on an equally-weighted combination of CCE and DWC. Statistical Techniques Our rst hypothesis is that statistically signicant differences exist among industries with respect to the
Research Findings
Average and Annual Working Capital Management Performance Working capital management component denitions and average values for the entire 1996 2000 period are given in Table 3. Across the nearly 1,000 rms in the survey, cash ow from operations, dened as cash ow from operations divided by sales and referred to as cash conversion efciency (CCE), averages 9.0 percent. Incorporating a 95 percent condence interval, CCE ranges from 5.6 percent to 12.4 percent. The days working capital (DWC), dened as the sum of receivables and inventories less payables divided by daily sales, averages 51.8 days and is very similar to the days that sales are outstanding (50.6), because the inventory
Table 3 Average Working Capital Scoreboard Variables Available Data Over Five Years
CFO Working Capital Measures Cash Conversion Efciency Days Working Capital Days sales outstanding Days payables outstanding 2000 10% 59 days 1999 9% 46 days 1998 10% 52 days 1997 10% 52 days 1996 6% 50 days
49 days
50 days
52 days
52 days
50 days
27 days
34 days
33 days
33 days
33 days
Inventory turns/year 10 times 12 times 11 times 11 times 11 times (days) (37 days) (30 days) (33 days) (33 days) (33 days) SOURCE: Annual Working Capital Surveys, CFO
13
Filbeck and Krueger
turnover rate (once every Table 4 32.0 days) is similar to Overall Working Capital Performance the number of days that By Industry with at Least an Average of Eight Companies Per Year payables are outstanding 1996 1999 (32.4 days). In all instances, the standard Industry Mean Overall CFO Standard Deviation of Range of Rankings deviation is relatively Ranking of Working Working Capital across All Firms and small, suggesting that Capital Performance Performance Years these working capital Petroleum 6 6 26 management variables Electric & Gas Utility 24 8 35 are consistent across Food Service 103 40 338 CFO reports. Telecommunications 122 240 882 The low standard Publishing 166 48 195 deviations reported in Pharmaceuticals 183 48 401 Table 2 are accentuated by the individual year Forest Products 186 66 273 values presented in Table Chemicals 193 75 309 3. As one might expect, Food 245 87 338 given a gross domestic Computers 247 97 367 product growth rate Beverage 255 238 857 range of only 5.6 percent Motor Vehicles 283 106 454 to 6.5 percent, there is Food & Drug Stores 287 61 243 relatively little difference Building Materials 296 121 502 in the CCE and DWC values. In 1996, CCE Electronics 306 141 499 was at a low of 6.0 Specialty Retailers 313 115 458 percent. Otherwise, the Health Care 365 144 623 CCE ratio was between Metal Products 399 147 564 9 and 10 percent. DWC Metals 448 117 413 reached a high of ftyWholesale 519 193 687 nine days in 2000, Furniture 531 234 904 mostly due to the slower General Merchandise Stores 554 154 577 inventory turnover in 2000. Otherwise, DWC Aerospace 624 220 720 values ranged from forty- Scientic Equipment 625 212 791 six to fty-two days. The Textiles 711 173 619 best year for working Apparel 720 177 728 capital management, as measured by a low days This table represents the average overall rank of a company within the stated industry (e.g., the average rank of the eight working capital gure companies included from the Petroleum industry during the sample period was 6). was 1999, when days payables outstanding reached a high of thirtyIndustry-based differences in overall working capital four days and inventory turnover reached a high of twelve management are presented in Table 4 for the twenty-six times per year (otherwise days payable outstanding ranged industries that had at least eight companies included in the between twenty-seven and thirty-three days, with inventory rankings each year. In the typical year, CFO magazine ranks turns between ten and eleven times per year). 970 companies during this period. Industries are listed in order of the mean overall CFO ranking of working capital Industry Rankings on Overall Working Capital performance. Since the best average ranking possible Management Performance for an eight-company industry is 4.5 (this assumes that CFO magazine provides an overall working capital the eight companies are ranked one through eight for the ranking for rms in its survey, using the following equation: entire survey), it is quite obvious that all rms in the Overall Ranking1 = (Highest overall CCE Company CCE) / (Highest overall CCE Lowest overall CCE) x petroleum industry must have (Lowest overall DWC Company DWC) / Lowest overall DWC Highest overall DWC) been receiving very high
14
Filbeck and Krueger
overall working capital management rankings. In fact, the petroleum industry is ranked rst in CCE and third in DWC (as illustrated in Table 5 and discussed later in this paper). Furthermore, the petroleum industry had the lowest standard deviation of working capital rankings and range of working capital rankings. The only other industry with a mean overall ranking less than 100 was the Electric & Gas Utility industry, which ranked second in CCE and fourth in DWC. The two industries with the worst working capital rankings were Textiles and Apparel. Textiles rank twenty-second in CCE and twenty-sixth in DWC. The apparel industry ranks twenty-third and twenty-fourth in the two working capital measures, respectively (also in Table 5). The second column of Table 4 exhibits the standard deviation in overall working capital performance rankings.
The industries with the greatest variation on the overall working capital performance measure, as measured by standard deviation, are the telecommunications industry and the beverage industry. If one only examines the extremes, the furniture industry is the industry with the greatest extremes in rank as it has at least one company whose rank varied from another rm in the same industry by 904 places. Variations in prot margins and turnover rates are worthy explanations for the wide disparity of rankings within the furniture industry. In general, the stability of rm rankings on WCM measures suggests that although a given level of current asset or current liability management impacts share price, one does not have to be overly concerned with changes in working capital management style.
Table 5 Average Industry Ranks of Working Capital Management Measures Across Components and Overall Ratings 1996 1999
Firm Aerospace Apparel Beverage Building Materials Chemicals Computer Electric/Gas Utilities Electronics Food Food Services Food Stores Forest Products Furniture General Merchandise Health Care Metals Metal Products Motor Vehicles Petroleum Pharmaceuticals Publishing Scientic Equipment Specialty Retailers Telecommunications Textiles Wholesale Trade Cash Conversion Efciency 20 23 12 14 5 8 2 9 11 10 26 6 18 24 17 19 15 16 1 3 4 13 21 7 22 25 Days of Working Capital 23 24 6 12 14 13 4 21 10 1 2 11 22 16 17 18 19 7 3 20 8 25 5 9 26 15 Days Sales Outstanding 19 11 7 10 20 16 8 22 4 2 1 5 17 6 23 9 15 13 25 21 14 26 3 24 28 12 Inventory Turnover 19 25 10 9 13 6 22 15 20 1 3 11 16 24 4 14 17 7 5 26 2 23 18 8 21 12 Days Payables Outstanding 19 21 6 8 1 13 10 7 16 26 25 12 20 14 23 9 4 2 5 11 24 18 17 3 22 15
15
Filbeck and Krueger
Industry Rankings Across Individual Working Capital Management Characteristics Table 5 breaks the overall working capital management rank in Table 4 into rankings of particular working capital measures (including the two components, CCE and DWC, which make up the overall rank). For instance, the petroleum industry, ranked rst for overall performance, only ranks rst in one of the ve specic working capital measures, CCE measure. In fact, as shown in the center column of Table 5, petroleums DSO performance is second worst among all industries. However, DSO is not included in the compilation of the overall rank. While Table 5 provides the relative rankings of industries across the ve working capital management measures, one may still wonder about the variation of these rankings over time. All of the instances wherein the standard deviation of rm rankings exceeded 5.0 are exhibited in Table 6. There were only eleven instances wherein the standard deviation exceeded 5.0. Both inventory turnover and days payables outstanding had a higher standard deviation in four instances. Only one industryTelecommunicationshad over two instances where the standard deviation of the industry ranking on a given working capital measure exceeded 5.0. One reason for this variation is the lack of stability in industry members, with over 60 percent of the rms in 1996 no longer in the study in 2000. Some of the other signicant changes include a dramatic drop in inventory turnover within the Petroleum industry and slower payment of accounts payable in the wholesale trade industry. In the other 93 instances (26 x 4 11), the variation in industry rankings for a working capital management variable is relatively stable. The number of days of working capital is relatively low in both the food services and food stores industries. Food stores, which are primarily cash-and-carry businesses, exhibit the shortest days sales outstanding ranking (with food services coming in second). However, food services have quicker inventory turnover, with the publishing
industry squeezing in between it and food stores. As one might imagine, food services, which like food stores tend to get payment upon purchase for merchandise, also need to make payments rapidly. In fact, these industries have the shortest days payables outstanding ranking, resulting in being at the bottom of the DPO column. Another factor hurting the performance of the food stores industry is its poor cash ow from operations per dollar of sales, resulting in it being ranked twenty-sixth in the Cash Conversion Efciency, the rst column of Table 5. Most industries were slower in collecting on sales than paying bills. In fact, only the food services, food stores, and specialty retailers had an average days payables outstanding value that exceeded their average days receivables outstanding. In addition, the beverage industry had a higher DPO than DSO value in three years, while the same relationship was true of the Apparel industry in only one instance. In all other eightyeight (26 x 4 16) instances, the industrys average DSO value was higher that year. Since CFO magazine only provides annual information, we are unable to assess the seasonal variation in WCM. All of the standard deviation data supplied illustrates the lack of much variation in WCM. Looking at the data, the most signicant trends existed in the Inventory Turnover measure, with the Beverage industry rising from eighteenth to eighth place, and the Telecommunications industry dropping from the second to twenty-rst position. Telecommunications also has a slower average collection period and quicker payment to suppliers, resulting in their DWC ranking dropping from rst to twenty-fth place. The only other trend in the data was the improvement (slowing) of payments to suppliers in the wholesale trade industry. Six industriesfood service, food and drug stores, forest products, petroleum, pharmaceuticals, and publishing rank in both the highest three and lowest three levels for at least one of their working capital performance rank measures. Table 5 illustrates that three industries (aerospace, building materials, and furniture) show the ve individual working capital performance Table 6 rankings are within six Instances Where the Standard Deviation of Rankings Exceed 5.0 places of each other. Of course, not having extremely different levels Cash Days of Days of performance across Days Sales Inventory Conversion Working Payables Outstanding Turnover individual working Efciency Capital Outstanding capital measures is not necessarily good. The Beverage 5.1 aerospace industry has Electric/Gas Utilities 6.1 the worst performance in Furniture 5.3 Days of Working Capital Health Care 6.6 (ranked twenty-third), Petroleum 7.2 but is only worthy of the Telecommunications 9.9 10.3 9.0 6.2 nineteenth ranking for Wholesale Trade 6.3 5.7 days sales outstanding, its best performance. 16
Filbeck and Krueger
Table 7 Analysis of Working Capital Management Overall and Across Components Working Capital Management Aspects Identied by CFO magazine 1996-1999
Industry Signicance Measure using ANOVA F-value Cash Conversion Consistency Industry Signicance 20.60** Days of Working Capital 21.54** Days Sales Outstanding 35.47** Inventory Turnover 22.12** Days Payables Outstanding 13.62** CFOs Overall Ranking 14.72**
Inter-Year Period Consistency Measured using Kendalls Coefcient of Concordancea Period Consistency
a
86.85**
87.35**
91.91**
87.64**
85.15**
83.24**
The signicance of Kendalls coefcient of concordance statistic (W) is measured using chi-square values, calculated as follows: x2 = Number of years (Number of industries - 1) W The critical value using Kendalls Coefcient of Concordance (alpha = 0.01) is 44.31. Signicance: * = 0.05; ** = 0.01
Statistical Signicance of Raw Numbers Table 4 and Table 5 report ordinal rankings of industries across working capital management variables. The ordinal rankings might be creating differences across industries that are, in reality, quite minute. Given the wide range of industry performance rankings, one might wonder whether there is a signicant difference in industry performance within individual aspects of working capital management. Table 7 shows the tests related to our two hypotheses. In the rst row (industry signicance), we nd support for our rst hypothesis that signicant differences exist between industries across time with respect to measures of working capital measures. The greatest differences occur in the days sales outstanding ranking, which has a statistically signicant ANOVA F-value of 35.47. Table 7 shows persistent statistical signicance, which suggests that there are signicant differences in the industry working capital management rankings. The second row in Table 7 (period consistency) shows the results related to our second hypothesis regarding the consistency of working capital measures within industries through time. This answers the question, are the rms cash conversion values consistent from period to period? Table 3 shows that despite the consistency in average values presented, there are signicant changes in individual rm values from year to year, based on the signicance of each of the values in the second row of Table 7. In other words, working capital measures for a given rm are not static, and signicant differences in these measures exist across time. With only four years of observations, the critical Kendalls Coefcient of Concordance value is 44.31. Yet, each of the Kendall Coefcient of Concordance values tends to be about twice this level. These results indicate that working capital measures vary across time. Taken together, our results in
Table 7 indicate that while working capital management ratios are changing over time for the rms sampled, these changes are consistent enough across industries to preserve the industry ordering across time.
Conclusions
The research presented here is based on the annual ratings of working capital management published in CFO magazine. Our ndings indicate a consistency in how industries stack up against each other over time with respect to the working capital measures. However, the working capital measures themselves are not static (i.e., averages of working capital measures across all rms change annually); our results indicate signicant movements across our entire sample over time. Our ndings are important because they provide insight to working capital performance across time, and on working capital management across industries. These changes may be in explained in part by macroeconomic factors. Changes in interest rates, rate of innovation, and competition are likely to impact working capital management. As interest rates rise, there would be less desire to make payments early, which would stretch accounts payable, accounts receivable, and cash accounts. The ramications of this study include the nding of distinct levels of WCM measures for different industries, which tend to be stable over time. Many factors help to explain this discovery. The improving economy during the period of the study may have resulted in improved turnover in some industries, while slowing turnover may have been a signal of troubles ahead. Our results should be interpreted cautiously. Our study takes places over a short time frame during a generally improving market. In addition, the survey suffers from survivorship bias only the top rms within
17
Filbeck and Krueger
each industry are ranked each year and the composition of those rms within the industry can change annually. Further research may take one of two lines. First, there could be a study of whether stock prices respond to CFO magazines publication of working capital management ratings. Second, there could be a study of which, if any, of the working capital management components relate to share price performance. Given our results, these studies need to take industry membership into consideration when estimating stock price reaction to working capital management performance.
About the Authors

Dr. Greg Filbeck serves as Senior Vice President of Schweser Study Program and Adjunct Professor of Research at the University of Wisconsin-La Crosse. He earned his doctorate in nance from the University of Kentucky. Prior to joining Schweser Study Program in 1999, he enjoyed ten years of fulltime university teaching experience from Miami University and the University of Toledo and has published over forty academic articles. greg.lbeck@schweser.com Dr. Thomas M. Krueger is a Professor of Finance at the University of Wisconsin-La Crosse. He earned his doctorate from the University of Kentucky. His teaching repertoire includes investments, corporate nance, and decision making. He is the past president of the Academy of Finance and is also the ECESP Internship Coordinator. His research includes fty journal articles, including the Super Bowl Stock Market Predictor and equity anomalies. Krueger.thom@uwlax.edu
Note
1. This ranking was not published in CFO magazine in 2002 or available at its Web site.
References
Corman, L. 1998. The 1998 working capital survey: Cash masters. CFO 14 (7):30-48. Farragher, E., R. Kleiman, and A. Sahu. 1999. Current capital investment practices. Engineering Economist 44 (2): 137-150. Fink, R. 2001. The 2001 working capital survey: Forget the oat? CFO 17 (9):54-64. Gilbert, E. and A. Reichert. 1995. The practice of nancial management among large United States corporations. Financial Practice and Education 5 (1): 16-23. Hadley, S. 2004. Making the business case: Supply chain management. Strategic Management (April):28-34. Hill, N., W. Sartoris, and D. Ferguson. 1984. Corporate credit and payables policies: Two surveys. Journal of Cash Management 559-576. Largay, J. and C. Stickney. 1980. Cash ows, ratio analysis and the W.T. Grant Company bankruptcy. Financial Analyst Journal 36 (4):51-54. Maness, T. and J. Zietlow. 2004. Short-term Financial Management. Australia: Southwestern Press. Mintz, S. 1999. The 1999 working capital survey: Dollars in the details. CFO 15 (7):55-68. Mintz, S. and C. Lazere. 1997. The 1997 working capital survey: Inside the corporate cash machine. CFO 13 (6):54-68. Myers, R. 2000. The 2000 working capital survey: Cash crop. CFO 16 (7):59-82. Scherr, F. 1996. Optimal trade credit limits. Financial Management 25 (1):71-85. Schwartz, R. 1974. An economic model of trade credit. Journal of Financial and Quantitative Analysis 9:643-657. Towne, J. 2002. Black inkSix Sigma archives, case study #5 Thibodaux Regional Medial Center. www.hfma.org/resource. Waxer, C. 2003. Six Sigma costs and savings. www.isixsigma.com/ library. Weinraub, H. and S. Visscher. 1998. Industry practice related to aggressive/conservative working capital policies. Journal of Financial and Strategic Decisions 11 (2):39-46. Weisel, J., N. Harm, and C. Bradley. 2003. The cash factor. Strategic Management (Sept.):29-33.
18
IBBEA Implementation and the Relative Protability of Small Banks

Srinivas Nippani, Texas A&M University-Commerce Kenneth M. Washer, Texas A&M University-Commerce
Abstract
The enactment of Riegle-Neal IBBEA in 1994 encouraged bank mergers and acquisitions. Empirical evidence indicates that large banks benefited from IBBEA enactment. However, there is little, if any, evidence of the impact of the act on small banks profitability relative to large banks. This study examines the impact of IBBEA on the performance of small banks in the period preceding and following IBBEA implementation. Evidence is presented that indicates the return on assets of small banks was significantly less than that of larger banks in the post-IBBEA period. This is contrary to the results of the pre-IBBEA period when small banks profitability was competitive with and in some cases even better than large banks profitability. It is concluded that the enactment of IBBEA has placed small banks at a competitive disadvantage which could eventually lead to their demise.
both groups of small banks significantly underperformed groups of larger banks following IBBEA implementation. The underperformance is tested using t-tests and a dummy variable regression that controls for general economic conditions and interest rate movements.
This study examines IBBEAs impact on the performance of small banks relative to larger banks.
Introduction
The enactment of the Riegle-Neal Interstate Banking and Branching Efficiency Act (IBBEA) in 1994 permitted bank holding companies to acquire banks in any state after September 30, 1995. This act gave banks a chance to reorganize/restructure in order to improve their efficiency and profitability. Industry groups and experts generally welcomed the act, but there were skeptics questioning its benefits to small banks. There was widespread discussion about the potential advantages and disadvantages of the act in both the business press and academic studies. Sufficient time has now elapsed which allows for an empirical examination of the evidence of bank performance following IBBEA enactment. This study examines IBBEAs impact on the performance of small banks relative to larger banks. The average return on assets (ROA hereafter) of banks with total assets equal to or less than $100 million and banks with total assets between $100 and $300 million is compared to average ROA of larger banks to see if significant underperformance is present in the post-IBBEA period. The results indicate that
Table 1 shows that bank consolidation began long before the enactment of IBBEA. In 1988 (beginning of sample for this study) there were 13,373 banks. By the end of 2002, the number of banks had fallen to 7,797. Over this entire period, category 1 banks (banks with total assets below $100 million) fell by 6,386 banks. All other larger bank categories increased with the exception of category 4, which fell slightly. Bank consolidation has been a trend for several decades. An interesting question is whether or not IBBEAs impact increased or perhaps changed the incentive for consolidation?
Literature Review
Empirical studies have shown that the passage of IBBEA led to statistically significant gains in the banking industry. Brook, Hendershott and Lee (1998) document a value gain of $85 billion for the industry. Carow and Heron (1998) report that IBBEAs passage had a positive wealth effect for large bank holding companies. Fraser, Hooton, Kolari and Reising (1997) examine the wealth effects of a decision by the Office of Thrift Supervision (OTS) to permit interstate branching for federally chartered savings and loan associations. They report that large savings and loan associations and commercial banks generally experienced significant positive wealth effects but little or no reaction was found for smaller depository institutions following key OTS 19
Nippani and Washer
Table 1 Number of Banks in Various Size Categories at Different Points in Time

Category Bank Size 1988.1 10,694 1,798 529 330 22 13,373 1995.3 6,891 2,062 621 352 37 9,963 2002.4 4,308 2,280 840 307 62 7,797
1 Banks with average assets under $100 million 2 Banks with average assets between $100 and $300 million 3 Banks with average assets between $300 million and $1 billion 4 Banks with average assets between $1 billion and $15 billion 5 Banks with average assets over $15 billion Total Source: Federal Reserve Bank of St. Louis
announcements in 1991 and 1992. The researchers argue that IBBEA benefits large institutions and thus accelerates the trend towards consolidation without necessarily compromising the viability of smaller institutions. Prasad (1997) comments that The community bank executives fear that the United States will consolidate its banking system to the point of having less than 400 banks in the country once large banks cross state lines with branches. In a recent study, Carow and Kane (2002) examine the value of relaxing long-standing regulatory restraints on banks over the period 1970-2000 and make the following comment, The evidence indicates that the new financial freedoms may have redistributed rather than created value. Event returns are positive for some sectors of the financial industry and negative for others. Based on their study, deregulation of the industry appears to be beneficial to only some sections of the industry. Nippani and Green (2002) show that bank performance improved in the post-IBBEA period, but when they controlled for general economic conditions and interest rate movements, the impact of IBBEA on bank performance appears to be insignificant. The evidence suggests that some sections of the industry, possibly large banks looking to merge and consolidate, stand to gain from the relaxation of geographic restrictions, and other sections of the industry, possibly small banks choosing not to merge and consolidate, stand to lose. There is evidence in the popular press that small banks were resisting the passage of IBBEA. Colorados state governor, over the vehement objections of small banks, vetoed a bill that would have barred national banks from branching into or out of the state (OHara 1995a). Several other articles in the popular press support the theory that small banks were against passage. For example: Proponents of opting out argue that the federal law will give big banks too much firepower to set pricing within communities and will create inequities in the industrys regulatory burden
The debate is perhaps most intense in Texas, home to more than 240 small banks and a host of superregionals The Texas Independent Bankers Association has managed to unite agricultural and small business groups under the banner Texas for preservation of hometown banking Meanwhile debate is building in several other states where opt-out bills have been introduced: Colorado, Kansas, Missouri, Montana, Nebraska, New Mexico and Oklahoma. (OHara 1995b) Other influential industry observers predicted that small banks would not be disadvantaged by IBBEA. J. Alfred Broaddus, Jr., President of the Federal Reserve Bank of Richmond argued, There is every reason to believe that smaller banks will not only survive but also prosper in a banking environment that provides expanded interstate opportunities(PR Newswire 1995). One could also argue that small banks would thrive as they serve a niche that bigger banks find less appealing.
Research Focus
The motivation for this study comes from academic studies and the popular press that argue both in favor of and against IBBEA in reference to its potential impact on small banks. These arguments about the future of small banks following the passage of IBBEA make it very interesting and necessary to examine the actual impact IBBEA had on small bank performance. This study differs from other studies in that it is the first to empirically analyze small bank performance relative to that of larger banks. The hypothesis is simply that small bank profitability in the post-IBBEA period is significantly below large bank profitability. The rest of the paper is organized as follows: Section 4 describes the data and methodology used in the study, Section 5 provides the empirical evidence, and conclusions are in Section 6.
20
Nippani and Washer
Data and Methodology

The purpose of this study is to examine small bank profitability relative to large bank profitability in both the preand post-IBBEA periods. The proxy for bank profitability is ROA. Return on equity (ROE) is also a common measure of profitability but it is not analyzed because it is simply ROA adjusted for financial leverage. It can be argued that financial leverage noise would interfere with comparisons of bank profitability. ROA is robust in that bank regulators look at this measure in evaluating bank performance, as it is part of the CAMELS rating. Madura (2003) states: Banks fail when their earnings become consistently negative. A commonly used profitability ratio to evaluate banks is ROA, defined as earnings after taxes divided by assets. In addition to assessing a banks earnings over time, it is also useful to compare the banks earnings with industry earnings. This allows for an evaluation of the bank relative to its competitors. The sample includes quarterly ROA for various bank groups for the period 1988.1 through 2002.4. The data set is divided into two groups with the separator being IBBEA implementation (1995.3). There are thirty-one quarters in the pre-IBBEA period and twenty-nine quarters in the post-IBBEA period. A study examining the performance of the average small bank with the performance of the average large bank reveals the relative performance of small banks in the post-IBBEA period. Quarterly ROA data for the following categories of banks is obtained from the Federal Reserve Bank of St. Louiss Internet website: 1. Banks with average assets under $100 million 2. Banks with average assets between $100 and $300 million 3. Banks with average assets between $300 million and $1 billion 4. Banks with average assets between $1 and $15 billion 5. Banks with average assets over $15 billion The St. Louis Fed acknowledges that the source of the information is the Federal Financial Institutions Examination Councils Reports of Condition and Income for All Insured U.S. Commercial Banks. In order to determine whether or not small banks underperformed larger banks, DROA is calculated by taking the difference in quarterly ROA of small banks (Categories 1 and 2 above) and larger banks (categories 2-5 above) and naming this variable DROA. The following seven DROAs are calculated: DROA1-2 = Category 1 ROA less category 2 ROA (1) DROA1-3 = Category 1 ROA less category 3 ROA (2) DROA1-4 = Category 1 ROA less category 4 ROA (3) DROA1-5 = Category 1 ROA less category 5 ROA (4) DROA2-3 = Category 2 ROA less category 3 ROA (5) DROA2-4 = Category 2 ROA less category 4 ROA (6) DROA2-5 = Category 2 ROA less category 5 ROA (7)
Category 1 banks ROA for 1988.1 is the average for all 10,694 banks in that category. As banks exit this category for various reasons (strong growth, merger, bankruptcy, etc.) they no longer impact the ROA calculation for this category. This could possibly bias the results. Obviously, if banks leaving this category have higher ROAs then the categorys ROA will suffer. However, banks could just as easily leave due to bad performance resulting in bankruptcy or merger. If this is the case, the categorys ROA will increase due to the laggards exit.
ROA is robust in that bank regulators look at this measure in evaluating bank performance, as it is part of the CAMELS rating
First, a t-test is used to determine whether DROA is negative and significant in the post-IBBEA period. This would indicate that small banks underperformed bigger banks postIBBEA. Second, a t-test is used to compare pre-IBBEA and post-IBBEA DROA values. If small banks performed poorly in the post-IBBEA period as compared with the pre-IBBEA period, one would expect the mean DROA to be significantly lower in the post-IBBEA period. A dummy variable regression is used to examine the impact of IBBEA on small bank performance. Consistent with the Nippani and Green (2002), changes in real GDP and the prime rate are included as independent variables in the regression equation. Controlling for these macro-economic variables allows the examination of changes in ROA holding other important variables constant. Banks tend to perform better when the economy is expanding and interest rates are falling. These two variables are obtained from the St. Louis Feds website. Regressions for all seven DROA variables are estimated using the following equation:
DROAn-m=B0+B1(IBEADUMMY)+B2(RGDP+B3(RPRIME) (8)
DROAn-m is the dependent variable where n takes a value of either 1 or 2, and m takes a value of 2-5 depending on the size category comparison. 0, 1, and 2 are coefficient estimates. The IBBEADUMMY takes a value of 0 for the pre-IBBEA period and a value of 1 for the post-IBBEA period. The hypothesis is that this variables coefficient estimate (1) will be negative and significant, thus indicating that ROA for smaller banks is significantly less than ROA for larger banks in the post-IBBEA period. RGDP is the percentage change in real GDP and is included in the regression equation to control for macro economic changes in the economy. PRIME is the prime rate and is included in the equation to control for general changes in interest rates. The t-test results and regression results are presented in the next section. 21
Nippani and Washer
Table 2 Results of t-tests for Comparison Between Pre- and Post-IBBEA Periods for Small Banks Versus Larger Banks
Variable Name Pre-IBBEA mean (n=31) Post-IBBEA mean (n=29) Difference in Means 0.101** 0.264** 0.403** 0.331**
Panel A: Banks with average assets $100 million versus larger bank classes DROA1-2 -0.0042 (0.101) -0.143** (0.049) DROA1-3 0.038* (0.105) -0.225** (0.090) DROA1-4 0.039 (0.227) -0.365** (0.186) DROA1-5 0.264** (0.254) -0.067 (0.192) Panel B: Banks with average assets between $100 - $300 million versus larger bank classes DROA2-3 0.080** (0.074) -0.082** (0.062) DROA2-4 0.081* (0.211) -0.222** (0.159) DROA2-5 0.306** (0.231) 0.076** (0.164) *Indicates t-value significant at 0.05 level. **Indicates t-value significant at 0.01 level.
0.162** 0.303** 0.230**
Column 1 shows the various DROA comparisons. Column 2 displays the pre-IBBEA DROA mean and standard deviation (in parenthesis). Column 3 shows the post-IBBEA DROA mean and standard deviation (in parenthesis). The difference between the pre and post-IBBEA period means is shown in column 4.
Empirical Evidence
The results of the t-tests comparing pre-IBBEA DROA with post-IBBEA DROA for the various size categories are given in Table 2 above. Panel A of Table 2 focuses on category 1 banks and compares their ROA performance to banks in categories 2-5. Column 2 shows that small banks significantly outperformed larger banks in two of the four cases during the pre-IBBEA period. In no case, however, did small banks significantly under perform larger banks in this period. In the post-IBBEA period (column 3), the situation is dramatically different as DROA values are all negative with three of four being statistically significant. The difference between DROA for pre and post-IBBEA periods is shown in column 4. These differences are all significant at the 0.01 level which provides validity to the contention that small banks were harmed by IBBEA. Panel B of Table 2 compares the ROA performance of category 2 banks with banks in categories 3-5. The results in this panel are virtually identical to panel A. Category 2 banks outperformed larger banks in the pre-IBBEA period, and DROA values in column 2 are all statistically significant. As expected, in the post-IBBEA period small banks significantly underperformed two of the three classes of larger banks. The difference in DROA values in the pre- and post-IBBEA periods is significant in all cases. The evidence in Table 2 indicates that 1) small bank performance was significantly worse than large bank performance in the post-IBBEA period and, 2) this under performance was in contrast to equal if not better relative performance in the pre-IBBEA period. The regression results are presented in Table 3. Panel A compares the ROA of category 1 banks with banks in categories 2-5. Panel B compares the ROA of category 2 banks with banks in categories 3-5. The coefficient estimates of IBBEADummy are negative and significant in all seven regressions. This estimate ranges from -0.36 to -0.11 and
indicates that ROA for category 1 and category 2 banks fell anywhere from -0.36 to -0.11 relative to the ROA of bigger banks in the post-IBBEA period. The two other independent variables in the equation are generally insignificant or are only marginally significant. Perhaps the most important evidence presented in Table 3 is that despite controlling for general economic conditions and interest rates, small bank performance has been inferior to large bank performance in the post-IBBEA period.
Conclusions
This study examines the performance of small banks both prior to and after geographic restrictions on interstate banking and branching were abolished. The findings indicate that small banks were competitive with, if not superior to, industry averages prior to the passage of IBBEA, but their performance deteriorated significantly in the period following the implementation of the act. This finding also lends support to articles in the popular press that predicted that small banks might face tougher business conditions following IBBEA implementation. This research contributes to the literature by showing that since the implementation of IBBEA, small bank profitability is significantly below that of larger banks. There are several potential long-term consequences for small banks. First, sub-par ROAs will negatively impact CAMELS ratings. Regulators use these ratings in determining the health of the institution and may impose additional monitoring costs on small banks, thus decreasing ROAs even more. Second, small banks that badly underperform may invest in riskier assets in hopes of increasing their returns and thus endangering the quality of their assets. This may lead to an increase in deposit insurance premiums and thereby create long run problems for regulators and taxpayers.
22
Nippani and Washer
Table 3 Results of Regression Analysis

Dependent Variable Adjusted R-square Intercept IBBEAdummy RGDP PRIME -0.01* 0 0.02 0.03
Panel A: Banks with average assets $100 million versus larger bank classes DROA1-2 0.34 0.07 -0.11** -0.01 DROA1-3 0.65 0.07 -0.27** 0.01 DROA1-4 0.52 -0.1 -0.36** -0.05 DROA1-5 0.39 0.02 -0.31** 0 Panel B: Banks with average assets between $100 - $300 million versus larger bank classes DROA2-3 0.61 0 -0.16** 0.02 DROA2-4 0.47 -0.17 -0.27** -0.04 DROA2-5 0.34 -0.06 -0.20** 0.01 *Indicates value significant at 0.05 level. ** Indicates value significant at 0.01 level.
0 .03* 0.04**
The dependent variable is the difference in average ROA (DROA) of banks in various size categories. IBBEAdummy is a dummy variable assigned a 0 in the quarters from 1988.1 through 1995.3, and a 1 in the quarters from 1995.4 through 2002.4. RGDP is the quarterly percentage change in real Gross Domestic Product. PRIME is the prime rate at the end of each quarter.
Nippani and Green (2002) report that overall bank performance improved in the post-IBBEA period as compared with the pre-IBBEA period, but when controlling for general economic conditions and interest rate movements, the impact of IBBEA on bank performance appears to be insignificant. This study extends Nippani and Greens (2002) work by focusing on bank size in the pre- and post-IBBEA periods.
PR Newswire. 1995. Federal Reserve Bank of Richmonds President Comments on Banking Industry in Cross Sections. P.R. Newswire, Financial News Section, June 2.
About the Authors
References
Brook, Y., Hendershott, R., & Lee, D. 1998. The gains from takeover deregulation: Evidence from the end of interstate banking restrictions. The Journal of Finance, 53(6):2185-2204. Carow, K. A., and Heron, R.A. 1998. The interstate banking and branching Efficiency Act of 1994: A wealth event for acquisition targets. Journal of Banking and Finance. 22:175-196. Carow, K.A., and Kane, E. J. 2002. Event-study evidence of the value of relaxing long-standing regulatory restraints on banks, 1970-2000. The Quarterly Review of Economics and Finance 42(3):439-463. Federal Reserve Bank of St. Louis, http://www.research.stlouisfed. org/fred2/. Fraser, D. R., Hooton, J. L., Kolari, J.W., and Reising, J. J. 1997. The wealth effects of interstate branching. Journal of Banking and Finance. 21(5):589-611. Madura, J. 2003. Financial markets and institutions. Sixth Edition, Thomson -South Western. Nippani, S., and Green, K. W. 2002. The banking industry after the Riegle-Neal Act: Re-structure and overall performance. The Quarterly Review of Economics and Finance. 42(5):901-909. OHara, T. 1995a. Governor slaps a veto on Colorado Opt-Out Bill. The American Banker. Community Banking section. (March 14):1. OHara, T. 1995b. Circling of Wagons to Fend off Branching. The American Banker. Community Banking Section (February 15):6. Prasad, R. 1997. Can technology even things up for community banks? Mid-American Journal of Business. 12(2):13-21.
Srinivas Nippani got his Ph.D. in finance from the University of Arkansas. He is currently an Assistant Professor with the Economics and Finance department at the Texas A&M UniversityCommerce. His work has been published in The Journal of Financial and Quantitative Analysis, the Quarterly Review of Economics and Finance and the Journal of Economics and Finance among others. sri_Nippani@tamu-commerce.edu Kenneth M. Washer earned his DBA in finance from Louisiana Tech University. He is currently an Assistant Professor at Texas A&M University-Commerce. His work has been published in the Journal of Banking and Finance, the Journal of Applied Business Research and Financial Counseling and Planning among others. Kenneth_washer@tamu-commerce.edu
23
Nippani and Washer
MBAA International
March 15-17, 2006

Palmer House Hilton-Chicago
"America's Best Conference Value"
Fifty states, twenty-ve countries and over forty years is powerful evidence
A collegial organization welcoming academics and practitioners interested in applying scholarly research to today's business challenges
Since 1978, MBAA International has returned to the Palmer House in Chicago for its annual meeting. The March meeting typically includes eight to nine hundred participants, representing approximately forty states and a number of foreign countries. Participants come from a variety of institutions, large and small, public and private, attend the three-day conference. From its rst meeting in 1965, MBAA International has acted as a coordinating body for autonomous groups in a variety of business-related academic areas. Now, there are over 700 members, in eleven divisions: accounting; case research; administration; human resources and industrial relations; information systems and quantitative methods; international business; legal studies; management; marketing; and operations management and entrepreneurship.
PapER SUbJECT AREaS:

Accounting Business and Health Administration Business Economics Business/Society/Government Entrepreneurship Finance Human Resources/Industrial Relations Information Systems/Quantitative Methods International Business Legal Studies in Business Management Marketing Operations Management Case Research
PROGRam aND REGISTRaTION INFORmaTION:

Executive Director MBAA: Dr. W. Jeff Clark Jennings A. Jones College of Business Middle Tennessee State University Murfreesboro, TN 37132-0001 Telephone: (615) 898-2838 jclark@mtsu.edu Program Chair: Dr. Susan Haugen Dept. of Accounting & Finance UW-Eau Claire Eau Claire, WI 54702-4004 shaugen@uwec.edu
42
24
nd
Annual Conference
For continuously updating information visit:
WWW.mbaaCONFERENCE.ORG
Modeling Internet Operations Using Initial Public Offerings

Sameer Prasad, University of Wisconsin-Whitewater David C. Porter, University of Wisconsin-Whitewater Linda Yu, University of Wisconsin-Whitewater
Abstract
In this research we test the generalizability of an existing model for classifying information-intensive services that can be globally disaggregated to Internet services. This categorization allows us to judge which types of Internet Initial Public Offerings (IPOs) are likely to have superior performance. Specically, we hypothesize that Internet rms with higher information intensity, lower physical presence and lower customer contact needs will have a greater probability of generating larger risk-adjusted returns. We test these hypotheses on 340 Internet IPOs and nd partial support for the model. In particular, Internet rms with high information intensity and low customer contact need yield superior performance. However, rms with low physical presence underperform in our sample.
the connection between the type of Internet operation and nancial performance.
the recent market correction has left
many investors, entrepreneurs and venture capitalists wondering if Internet commerce is a viable medium to exchange transactions. To provide this empirical connection, we use the literature in information technology and service operations to locate an appropriate model. Although we nd several models with some applicability, we use the Apte and Mason (1995) model of disaggregation of information-intensive services owing to its relevance, support in the literature, and testability. The remainder of the paper is organized as follows. Section 2 contains the literature review and hypotheses, Section 3 the variable denitions, Section 4 the data and methodology, Section 5 the results and Section 6 concludes.
Introduction
E-commerce allows organizations to facilitate relationships with customers, buyers and suppliers. Some organizations expect this technology to reduce costs, improve quality and reduce lead times. Others hope e-commerce will serve as the basis for entirely new business models. One of the main portals for e-commerce investment is the Internet. The Internet provides a common platform not only for a companys internal activities, but also for connections with operations throughout the entire supply chain (Vakharia 2002). To generate sufcient capital for e-commerce investment, new rms often consider an Initial Public Offering (IPO). The late 1990s produced large numbers of new e-commerce ventures, many of which issued IPOs. Unfortunately, many of these new ventures had poor business models with respect to the Internet and although there was an initial run up in stock prices, the recent market correction has left many investors, entrepreneurs and venture capitalists wondering if Internet commerce is a viable medium to exchange transactions. If managers and investors are to appropriately channel investment funds toward Internet operations, one key to appropriate decisions is understanding which business models are more likely to benet from the technology. Unfortunately, the current literature provides little empirical evidence on
Literature Review and Hypotheses
The extant literature contains both theoretical and descriptive models in the area of Internet use (Matthew 1998; Rai, Ravichandran and Samaddar 1998; Lin 1999). Other researchers have explored the connections of e-commerce with a whole host of issues including trust (Kanawattanachai and Yoo 2002; Pavlou 2002), strategy (Lee 2001; Sweet 2001), and customer needs (Chen and Dubinsky 2003). However, these models do not indicate which business types would provide superior performance using the Internet. Extensive research exists on estimating the value of traditional IT investments (Gurbaxani and Whang 1991; West and Courtney 1993). On the other hand, the Internet literature indicates that the web is expected to provide a mechanism for cost reduction (Keeney 1999; Harris and Katz 1991), quality improvement (Finch 1999; Keeney 1999), and lead-time minimization (McKnight and Bailey 1997; Keeney 1999) but several of these models are based on internal analysis that is generally not available to investors when examining rms from external sources.
25
Prasad, Porter, and Yu
In this research, rather than attempting to estimate the cost/benet equations for the Internet ventures, our interest is in linking industry characteristics with nancial performance as measured by stock market returns. We assume that cost reduction, increased quality and reduced lead-times should lead to improved nancial performance of a company but the amount of such improvement could vary by industry classication. The literature on e-commerce provides some clues on how to categorize rms. For example, researchers have examined the economic implications of agent technologies and e-commerce (Vulkan 1999). E-commerce requires that agents be present. Agents are computer systems that enable autonomous operations (Vulkan 1999) and can yield results only when the activities can be disaggregated. In other words, the potential of e-commerce lies in the disaggregation of value chain activities in a manner that transcends both organizational and geographic boundaries (Apte and Mason 1995). Advances in IT such as the Internet allow information intensive activities to be dispersed out of the ofce (Apte and Mason 1995). Thus, it is essential that any model used to gauge the effectiveness of e-commerce organizations captures the potential to disaggregate activities. The same critical aspect of analysis is also found in the international information systems literature, where distances require a greater degree of disaggregation of components similar to the Internet. Three primary areas relevant to disaggregation are examined in the literature: customer contact, information intensity, and physical presence. The area of customer contact has been recognized as an important construct (Schmenner 1986; Haywood-Farmer 1988; Wemmerlov 1990). Considerable research (e.g. Karmarkar and Pitbladdo 1995; Soteriou and Chase 1998), also provides empirical support for customer contact theory. In addition, the academic literature related to CRM and online transactions indicates that the organizational network is focused around customers, not suppliers (Achrol and Kotler 1999). Achrol and Kotler (1999) conclude that at the core of this organizational network is a collection and dissemination of information about customers and the quality of knowledge represents the primary source of power. Other literature also points to the signicance of information within organizations. Constructs of information richness/intensity are included in several models including Porter and Miller (1985), Harris and Katz (1991), and Apte and Mason (1995). Finally, the role of physical ows is widely studied in the operations management and industrial engineering literature. Apte and Mason (1995) suggest that time and motion studies are the primary method used to measure and improve upon the physical processes but Lee and Whang (2001) extend the role of physical ows to e-fulllment efciencies. In summary, substantial support exists for each of the three disaggregation constructs, information attributes, customer contact need and physical presence. Hence, we propose three testable hypotheses of Internet rm performance: 26
H1: There is no signicant difference in performance between Internet rms with low versus high information intensity. H2: There is no signicant difference in performance between Internet rms that require low versus high customer contact. H3: There is no signicant difference in performance between Internet rms that require low versus high physical presence. To test the three hypotheses, we dene each construct as well as identify superior performance. These denitions are developed in the following section.
Classifying Internet Ventures and Performance
The Apte and Mason (1995) model provides a classication system able to identify industries that can be disaggregated using the three constructs. Their model identies the propensity to globally distribute different types of jobs based upon the three constructs. The nature of globally dispersing services implies a distance between customers and the organization. In this research, we believe that this concept of distance is appropriately applicable to the Internet. Measuring Information Attributes The degree of information content in a business is dened by both Apte and Mason (1995) and Porter and Miller (1985). Apte and Mason dene information intensity as the ratio of time spent in dealing with information in an activity to the total time spent in that activity. Porter and Miller suggest information can be measured along two dimensions: the information content of the product and the information intensity of the value chain. If we apply this analysis and use the same categorization scheme of Apte and Mason, (e.g. categorize each of these dimensions as either low, medium or high), the combination results in ve descriptive categories as indicated in Figure 1. Category one is dened as low information content and low information intensity of the value chain; category two, low information content and medium information intensity of the value chain, or medium information content and low information intensity of the
Figure 1 Information Attributes Grid
High Information Content Medium Low 2 1 Low
4 3 2 Medium
5 4 High
Information Intensity
value chain; category three, medium information content and medium information intensity of the value chain; category four, medium information content and high information intensity of the value chain, or high information content and medium information intensity of the value chain; and category ve, high information content and high information intensity of the value chain. As an example, we would classify a company like E-toys as category onelow in information intensity since the products contain low amounts of information and relatively little information exists within the value chain. Measuring Customer Contact Need Apte and Mason (1995) dene customer contact need as the interface between the customer and the service provider. Borrowing from Chase (1981), the construct is dened along two dimensions: in-person contact and symbolic contact. The level of contact need is a function of the inperson contact desired by the customer (internal/external), buyer, or supplier, and the degree to which information can be symbolically represented. Consistent with these previous studies, we dene the in-person contact dimension as mutual conding and trust. We partition it into three levels: low, medium and high. Symbolic contact is when the main purpose of a customers presence is to exchange the information necessary for service creation and consumption, and can be easily represented (Apte and Mason 1995). This dimension is partitioned into three different categories resulting in total ve descriptive categories as indicated in Figure 2. Category one is dened as high in-person and low in symbolic; category two, medium in-person and low in symbolic, or high in-person and medium in symbolic; category three, low, medium, or high for both; category four, low in-person and medium symbolic, or medium in-person and high in symbolic; and category ve, low in-person and high symbolic. Consider the following examples. In some operations such as purchasing shares, the need for in-person contact is low, whereas in the teaching environment desire for contact can be high. Symbolic representation is the degree to which information in a business could be represented appropriately and transferred over the Internet. For instance, nancial data and execution of orders would be relatively easy to symbolically represent, reducing the need for actual contact. Taste
Figure 2 Customer Contact Grid
and smell characteristics of products in a bakery, on the other hand, would be difcult to represent symbolically. Measuring Physical Presence Need Porter and Miller (1985) suggest that almost all value chain activities have a physical task/component associated with them. Physical items need to be transported, sorted or perhaps processed/assembled. However, it is important to remember that even manufacturers such as General Motors have substantial information necessary to enable value added operations. Thus, Apte and Mason (1995) dene the physical presence need as the ratio of time spent in physical actions to the total time spent in a service activity. Once again, this construct is partitioned into the three categories: a low amount of physical presence is classied at category three and a high amount of physical presence is classied at category one as indicated in Figure 3.
Figure 3 Need for Physical Presence Scale
1 High
2 Medium Physical Presence
3 Low
In-person Contact
High Medium Low
1 2 3 Low
2 3 4 Medium
3 4 5 High
Symbolic Contact
In summary, we use the denitions of information attributes, customer contact need and physical presence developed by Apte and Mason (1995) to categorize Internet rms into a 5x5x3 matrix: ve categories for information attributes, ve categories for customer contact need and three categories for physical presence. This matrix is then used to determine if there are signicant differences in the Internet IPO performance across each of the three constructs. For example drugstore.com, Inc. is an online retailer that sells name-brand and private-label health and beauty products and prescription drugs. It also sends rell reminders via e-mail and offers detailed information about drugs and health issues. This IPO was classied as having low information content and low information attributes of the value chain, high in person and low in symbolic, and high in physical presence. Hence, for the information attributes rating it received 1, for customer contact 1, and physical presence 1. MapQuest.com, Incs online products and services are geared for people who want to go someplace. The company (formerly GeoSystems Global Corporation) got its start as the cartographic unit of R. R. Donnelley & Sons that supplies maps and proximity information. We classied this IPO as high in information content and high information intensity of the value chain, low in person and high in symbolic, and low in physical presence. Thus, MapQuest.com, Inc was rated 5 for information attributes, 5 for customer contact and 3 for physical presence.
27
Measuring Internet IPO Performance The nance literature provides several models for measuring performance. In most of these models there is a two dimensional trade-off between risk and returnas risk increases, expected returns also increasetherefore it is not possible to measure performance without measuring both risk and return. The historical existence of the risk-return trade-off is demonstrated methodically with the annual publication of Ibbotson Associates yearbook Stocks, Bonds, Bill, and Inationthe lowest risk bills average the lowest returns followed by bonds and then by stocks. Small rms have the highest risk and have the highest average returns over the period of the study. The fact that bills have lower returns than small rms is a meaningless comparison until the returns are adjusted for the differences in risk between the two groups of securities. The Ibbotson yearbook uses existing security prices to generate risk and return measurements. Risk and return measurement in the IPO market is much more difcult since IPOs, by their nature, do not have historical information available to gauge risk classications and IPOs have known return measurement anomalies. There is substantial evidence demonstrating that IPO markets contain two anomalous returns periods, initial underpricing and long-term overpricing. Underpricing implies that the initial offering price is lower than market expectations and there is usually a dramatic price increase in the rst day of IPO trading. In a review paper by Ritter and Welch (2002) about IPO activities, we see IPO shares trade on average at 18.8 percent above the original issue price during 1980 to 2001. On the other hand, the long-run performance of these hot IPOs consistently underperforms seasoned companies by an average of 23.4 percent for the same time period. For several decades, the IPO literature has been virtually unanimous in its conclusion that IPOs are underpriced in the short term (Ibbotson 1975; Ibbotson, Sindelar and Ritter 1988; Levis 1993, Bradley and Jordan 2002) and that most of the markets response to the underpricing occurs on the rst few trades of the rst trading day after the IPO issue (Barry and Jennings 1993). Critics have charged that the rst day run-ups benet Wall Street rms and their best customers but hurt individual investors who rarely have an opportunity to purchase IPOs at the offering price. The recent advisory committee formed by the NYSE and NASD at the request of SEC chairman Harvey Pitt demonstrates the concern over this long standing issue. Several theories have been forwarded to explain the initial underpricing including winners curse (Rock 1986), costly information acquisition (Benveniste and Spindt 1989), information asymmetries between issuers and their investment bankers (Baron and Holmstrom 1980), avoidance of legal liability (Tinic 1988), and signaling (Grinblatt and Chuan 1989), but no theory adequately explains the initial IPO underpricing. The lack of a consensus explanation for the anomaly has led many researchers (e.g. Kunz and 28
Aggarwal 1994; Aggarwal and Rivoli 1990) to remove early trading from their research. Consistent with this analysis, we follow the same procedure for generating initial IPO prices. More complex is the long-run (1 to 5 year) overpricing that occurs after the IPOs are listed. Again, although several theories attempt to explain the long-run overpricing (Schultz 2003; Ritter 1991; Reilly 1977), none adequately explains the IPO long-run performance and there is no consensus on when the long-run overpricing occurs. Further, the current theories do not differentiate which industries may under- or over-perform across a specied time frame. During the rst 180 days (6 months) following an IPO, company employees, directors, venture capitalists and other big shareholders are prohibited from selling their shares. This common practice is called lockup period. As soon as the lockup period expires, these big shareholders usually sell and cause stock price to fall in the secondary market (Aggarwal, Krigman, and Womack, 2002). Since the IPOs are underpriced in the short term and overpriced in the longrun, the choice of a return period for performance measurement is concerning. For purposes of this paper, we use a six month return period which connes our study within the lockup period. Although the choice of six months is arbitrary, it is chosen in an attempt to balance the two anomalies and to avoid the selling pressure effect once the lockup period expires. Statistical tests computed later in the paper imply that the six month choice adequately balances the two anomalies.
Data and Methodology
Our initial sample includes 2,967 IPOs over the period January 1, 1995 to December 31, 2002 covering the pre- and post-Internet bubble period. These IPOs are screened for Internet related businesses using Internet related denitions (e.g. Internet, Internet Services, Internet Development (Services), Internet Healthcare) from IPO.com, IPO Maven and Media General. IPOs not listed as Internet related by any denition are deleted from the sample resulting in 391 IPOs. These IPOs are then screened for complete data over the test period resulting in a nal sample of 340 IPOs. Complete data are dened as having both stock price information and the company information necessary to categorize the rm within the three constructs. To measure market adjusted returns, we follow the methodology of Aggarwal and Rivoli (1990): ARit = Pit - Pi1 Pi1
Mit - Mi1 Mi1
where ARit is the market-adjusted return of stock i at time t, Pit is the price of stock i at time t, Pi1 is the price of stock i at the end of the rst day of trading of the IPO, and M is similarly dened for an appropriate market index. Since all of our IPOs are listed on the Nasdaq, we use the Nasdaq as the
appropriate market index. The earlier discussion and capital market theory dictates the use of a broad market index rather than a portfolio of IPOs since investors are assumed to hold diversied portfolios and are interested in how a stock affects a broad based portfolio rather than its performance within a small group. As discussed in the previous section, we used a six month return period. Each rm is assessed within the three constructs. Classications are based on descriptions of company strategy, marketing plan, products and operational setups. Descriptive statistics are generated including: 1) a frequency plot of the number of IPOs issued by year, 2) distribution of rms by the various levels within the three constructs, and 3) distribution by type of business. To test the internal validity of the classication process, we employ an external auditor to re-rate each rm. For each construct, we compute the difference in measurement between the raters and test for signicant differences. We found no signicant difference in construct classications and the inter-rater reliability at 0.88. A larger team of raters would improve the classication reliability but the similarity of the classications between the independent raters in this study suggests that
managers could have internet projects or ventures ranked by small teams and at low cost, and yet achieve high inter-rater consistency. Four additional statistical tests are used to determine performance differences. First, a t-test for difference of means is used to determine if Internet rms, on average, have greater returns than the Nasdaq over the study period. Second, a ridge regression was employed to examine the possible inuence of multicollinearity among the parameter estimates. Third, a linear regression is used to test the validity of the Internet model using net return as the dependent variable with information attributes, physical presence and customer contact as independent variables. Finally, we also ran a regression based on one year net returns relative to information attributes, physical presence and customer contact classications.
Results
Descriptive Statistics Figure 4 shows our sample of Internet IPOs is more heavily information intensive, implying that the interface
Figure 4 Frequency Distribution within Three Dimensions*

Information Attributes Customer Contact Need
Physical Presence
Category 1 (56)
Category 2 (11) Category 3 (25) Category 4 (31)
Category 1 (35) Category 2 (26) Category 5 (156) Category 3 (86)

Category 3 (165) Category 1 (84)
Category 5 (217)
Category 2 (91)
* N = 340 Construct Information Attributes Category 1 (Description) Low information content & low information intensity Category 2 (Description) Low information content & medium information intensity, or medium information content & low information intensity Medium in person & low in symbolic, or high in person & medium in symbolic Medium
Category 4 (37)
Category 3 (Description) Medium information content & medium information intensity Category 4 (Description) Medium information content & high information intensity, or high information content & medium information intensity Low in person & medium symbolic, or medium in person & high in symbolic Category 5 (Description) High information content & high information intensity
Customer Contact
High in-person & low in symbolic
Low, medium, or high for both
Low in person & high symbolic
Physical Presence
High
Low
29
Figure 5 Frequency Distribution of the Number of Internet IPOs Issued by year.
200 Count 150 100 50 0 1995 1996 1997 1998 1999 2000 2001 2002
Year
between customer and service provider can be easier to support and requires less physical movement of goods and people. Figure 5 presents the IPO distribution by year and indicates that although Internet IPOs are offered in each of the eight-year sample, the majority were issued in 19982000. Table 1 provides frequency counts of business lines for the Internet IPOs and demonstrates the wide distribution across business type, with software being the most common followed by content and retail operations. Inferential Statistics The IPO mean net return over the Nasdaq was 2.15 percent with a standard deviation of 132 percent. Given the 339 degrees of freedom, the resulting t-value of 0.0163 is not signicant and implies that our choice of a six-month returns period adequately balances the underpricing and overpricing concerns discussed earlier. Also, as expected our regression analysis over a one-year span yielded a minimal adjusted r-squared value of .023, further reinforcing our choice of using a six-month span. A ridge regression trace of the estimated standardized regression coefcients (s) is presented in Figure 6. As it is apparent in Figure 6, the signs of the s do not change throughout the entire spectrum of the biasing constants K. Hence, we can safely ascertain that the multicollinearity is not severe enough to distort the parameter estimates (Neter, Wasserman and Kunter, 1985). Thus, the use of the Ordinary Least Squares (OLS) is appropriate. Running the three independent variables (information attributes, physical presence and customer contact) with the dependent variable (market adjusted abnormal return) in an OLS regression yielded an R value of 31%, implying the model has some explanatory power. Table 2 shows the regression coefcients, standard errors, and t-statistics for the three independent variables. All three coefcients are signicant (p .05). Hence, we can reject all three null hypotheses that there is no difference 30
in performance between rms with low versus high values for each construct. Firms with greater information intensity exhibited above average market-adjusted performance ( = 38.62, p 0.01). The positive coefcient indicates that Internet rms with relatively high information intensity will on average yield 38.6 percent market-adjusted performance. This nding is consistent with our earlier discussion since rms with signicant information intensive value added activities should be more successful using the Internet over the long-term than rms such as retail outlets with much less information intensity. Inconsistent with expectations, rms with lower physical presence need in their operations, yield lower market-adjusted returns ( = 51.06, p= 0.01). The negative coefcient suggests that organizations with lower physical presence will yield 51.1 percent worse market-adjusted performance relative to other Internet IPOs in our sample. One possible explanation is that physical operations usually revolve around manufactured items where value has been long established. For services, value is more difcult to ascertain. For example, a manufactured item provides value for years, whereas a service could be instantly perishable. Another possible explanation for the lower yield for low physical presence rms could be that the startup costs/barriers are low for these rms causing a fast proliferation of competitors.
Table 1 Frequency Distribution by Type of Business

Administration/Transactions Auctions Collaborations Communities Consulting Content E-business Education E-mail Financial/Insurance Games/Entertainment Hardware Hardware/Software integration HRM Infrastructure ISP Links Marketing Mix Monitoring Music Others Portal Real Estate Retail Search Engines Software Travel Web-hosting 9 11 1 2 13 29 12 6 5 21 3 12 11 4 19 15 5 19 5 7 7 5 9 3 29 3 64 5 6
Figure 6 Ridge Regression Trace for the Three Constructs

0.5 0.4 0.3 0.2 0.1 0 -0.1 -0.2 -0.3 -0.4 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 K 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95
information
customer
physical
Finally, the customer contact variable also yielded statistically signicant results ( = 8.98, p 0.05) indicating that Internet ventures, where the desire for in-person customer contact is low and the information can be symbolically represented with ease, will yield higher market-adjusted performance. The results imply that rms that can more easily transcend the interface between customer and service provider should yield 9 percent better performance.
Conclusion
This research can provide direction for managers, investors, IT entrepreneurs and venture capitalists by suggesting the type of startups that would yield above average marketadjusted performance and perhaps the Internet models that might be more successful. Specically, we nd that IPO rms having high information attributes and low customer contact need signicantly outperform other Internet IPOs after adjusting for market-related risk. Contrary to expectations, rms with low physical presence need have signicantly worse market-adjusted performance relative to other Internet IPOs in our sample. The results of this research are valuable to both academics and practitioners. Entrepreneurs of Internet operations
Table 2 Linear Regression Between Net Returns and Internet Dimensions
Independent Variables Information intensity Customer contact Physical presence Unstandardized () 38.62 8.98 51.06 Coefcients Standard Error 6.80 5.23 12.92 tvalue 5.679 1.714 3.95 Signicance p 0.01 p 0.05 p 0.01
can use this information to determine what type of business should be moved to the Internet. In addition, venture capitalists, investors and managers can apply this model to channel investments into projects more likely to generate superior performance. An interesting nding of our research was that those rms with a greater physical presence tended to have on average higher market adjusted returns. Hence, it might be possible to extrapolate that customers might prefer brick and click business relative to only on-line operations. Making such generalizations could be risky, given that the analysis was based only upon online rms and that market adjusted returns (not net returns) were used to measure performance. These research ndings are also of value in linking the literature in service operations and global information systems to Internet based operations and show that a model like the Apte and Mason (1995) one of global disaggregation can be generalized to other mediums. It is important to understand some of the assumptions that might limit the generalizability of our ndings. Since we tested the performance of Internet IPOs in the United States, the results may not be directly applicable to traditional brick and mortar operations moving into e-commerce venues. In addition, the conclusions from this analysis are only valid to Internet IPOs listing in the United States. Finally, as technology changes rapidly overtime, some variables such as the symbolic representation within the customer contact might be dramatically affected. These issues naturally lead to the possibility of future research. Our model seeks to identify market performance of Internet IPOs by business type. The overall model is able to explain 31 percent of the returns behavior. Additional variables might increase the models explanatory power. For example, a large part of a rms performance is also based on the its characteristics and management. Research on the interactions between management policies and the three constructs
31
would be useful. In addition, it would be worthwhile to replicate this study on other high-technology stock markets in other developed regions (Canada, Europe, and Japan) as well as in developing countries (India and Taiwan). Researchers might also explore the role of the three constructs utilized in this study and compare the performance between only online organizations with those that are brick and click. Finally, it would be benecial to gauge how changes in technology interact with information attributes, customer contact, and physical presence as measured by nancial success.
References
Achrol, R.S. and P. Kotler. 1999. Marketing in the network economy. Journal of Marketing 63(special issue):146-163. Aggarwal, R., L. Krigman, and K. Womack. 2002. Strategic IPO underpricing, information momentum, and lockup expiration selling. Journal of Financial Economics 66(1):105-137. Aggarwal, R., and P. Rivoli. 1990. Fads in the initial public offering market? Financial Management 19(4): 45-57. Apte, U.M., and R.O. Mason. 1995. Global disaggregation of information-intensive services. Management Science 41(7): 1250-1262. Baron D., and B. Holmstrom. 1980. The investment banking contract for new issues under asymmetric information: delegation and the incentive problem. Journal of Finance 35(5): 11151138. Barry, C., and R. Jennings. 1993. The opening price performance of initial public offerings of common stock. Financial Management 22(1):54-64. Benveniste, L., and P. Spindt. 1989. How investment bankers determine the offer price and allocation of new issues. Journal of Financial Economics 24(2):343-361. Bradley, D., and B. Jordan. 2002. Partial adjustment to public information and IPO underpricing. Journal of Financial and Quantitative Analysis 37(4):595-616. Chase, R.B. 1981. The customer contact approach to services: theoretical bases and practical extensions. Operations Research 21(4):698-705. Chen, Z. and A.J. Dubinsky. 2003. A conceptual model of perceived customer value in E-commerce: A preliminary investigation. Psychology Marketing 20(4):323-348. Finch, B. 1999. Internet discussions as a source for consumer product customer involvement and quality information: an exploratory study. Journal of Operations Management 17(5): 535-556. Grinblatt, M. and H. Chuan. 1989. Signaling and the pricing of new issues. Journal of Finance 44(2): 393-420. Gurbaxani, V. and S. Whang. 1991. The impact of information systems on organizations and markets. Communications of the ACM 34(1):59-73. Harris, S.E. an L.L. Katz. 1991. Predicting organizational performance using information technology managerial cost ratios. Proceeding of the Twenty-Second Annual Hawaii International Conference on Systems Sciences 4,197-204. Haywood-Farmer, J. 1988. A conceptual model of service quality. International Journal of Operations and Production Management 8(6):9-29.
Ibbotson, R. 1975. Price performance of common stock new issues. Journal of Financial Economics 2(2):235-272. Ibbotson, R., J., Sindelar and J. Ritter. 1988. Initial public offerings. Journal of Applied Corporate Finance 1(2):37-45. Kanawattanachai, P. and Y. Yoo. 2002. Dynamic nature of trust in virtual teams. Journal of Strategic Information Systems 11(34):187-213. Karmarkar, U., and R. Pitbladdo. 1995. Service markets and competition. Journal of Operations Management 12(3-4): 397-411. Keeney, R. 1999. The value of Internet commerce to the customer. Management Science 45(4): 533-542. Kunz, R., and R. Aggarwal. 1994. Why initial public offerings are underpriced: evidence from Switzerland. Journal of Banking and Finance 18(4):705-723. Lee, C-S. 2001. An analytical framework for evaluating e-commerce business models and strategies. Internet Research: Electronic Networking Applications and Policy 11(4):349-359. Lee, H.L. and S. Whang. 2001. Winning the last mile. Sloan Management Review 42(4): 54-62. Levis, M. 1993. The long-run performance of initial public offerings: the UK experience 1980-1988. Financial Management 22(1):28-41. Lin, C.A. 1999. Online-service adoption likelihood. Journal of Advertising Research 39(2):79-89. Matthew, L. 1998. Internet-based nancial EDI: Toward a theory of its organizational adoption. Computer Networks 30(1618):1579-1588. McKnight, L.W. and L.P. Bailey. 1997. Internet Economics Cambridge, MA: The MIT Press. Neter, J., W. Wasserman, and M.H. Kunter. 1985. Applied Linear Statistical Models Homewood, IL: Irvin. Pavlou. P.A. 2002. Institution-based trust in interorganizational exchange relationships: the role of online B2B marketplaces on trust formation. Journal of Strategic Information Systems 11(5-6):215-243. Porter, M. and V. Miller. 1985. How information gives you competitive advantage. Harvard Business Review 63(4):149-161. Rai, A., T. Ravichandran and S. Samaddar. 1998. How to anticipate the Internets global diffusion. Communications of the ACM 41(10):97-106. Reilly, F. 1977. New issues revisited. Financial Management 6(4):28-42. Ritter, J. 1991. The long-run performance of initial public offerings. Journal of Finance 46(2):3-27. Ritter, J.R. and I. Welch. 2002. A review of IPO activity, pricing, and allocation. Journal of Finance 57(4):1795-1828. Rock, K. 1986. Why new issues are underpriced. Journal of Financial Economics 15(1/2):187-212. Schmenner, R.W. 1986. How can service businesses survive and prosper? Sloan Management Review 28(2):21-32. Schultz, P. 2003. Pseudo market timing and the long-run underperformance of IPOs. Journal of Finance 58(2):483-518. Soteriou, A. and R.B. Chase. 1998. Linking the customer contact model to service quality. Journal of Operations Management 16(4):495-508. Sweet, P. 2001. Strategic value conguration logics and the new economy: a service economy revolution? International Journal of Service Industry Management 12(1):70-84. Tinic, S. 1988. Anatomy of initial public offerings of common stock. Journal of Finance 43(4):789-822. Vakharia A.J. 2002. E-business and supply chain management. Decision Sciences 33(4):495-504.
32
Vulkan, N.1999. Economic implications of agent technology and e-commerce. The Economic Journal 109(453):67-90. Wemmerlov, U. 1990. A taxonomy for service processes and its implications for system design. International Journal of Service Industry Management 1(3):13-27. West, L.A. and J.F. Courtney. 1993. The information problems in organizations. Decision Sciences 24(2):229-251.
MIDWEST
FINANCE
ASSOCIATION
TH
March 23-25, 2006 Chicago Marriott, Chicago, IL
About the Authors

Sameer Prasad is Professor of Management at the University of WisconsinWhitewater. His research interests lie in the areas of supply chain management, e-commerce, and global operations management. prasads@uww.edu David Porter is Professor of Finance at the University of WisconsinWhitewater. His main area of research is market microstructure. porterd@uww.edu Linda Q. Yu is Assistant Professor of Finance at the University of WisconsinWhitewater. Her research interests lie in the areas of Treasury Ination-Indexed Securities, Initial Public Offering, and market microstructure. yuq@uww.edu
55
ANNUAL MEETING
Members and friends of the MFA are invited to submit papers to be considered for presentation at the annual meeting of the Midwest Finance Association. Papers on any topic related to nancial economics, nancial planning education will be considered.
Paper Submission Deadline: September 28, 2005

Each paper will be reviewed by at least two members of the program committee, the Track Chair and the Program Chair. Authors will be notied of the selection desicion in November 2005. Accepted papers are eligible to be published in the proceedings.
Questions? - contact:
Larry J. Prather 2006 First Vice President and Program Chair Email: prather@etsu.edu
(No manuscripts please)
Submission Information:
All information regarding electronic submission, program committee, awards, special sessions and registration is available at the ofcial web site of the 2006 meeting:
www.mafa-2006.org
33
NCSM 2006
National Conference in Sales Management
March 1-4, 2006 Minneapolis, Minnesota

Special panels will bring industry practitioners and educators together to discuss training and teaching techniques and topics to improve professional selling and sales management classes.
Professional Sales and Sales Management Practices for the 21st Century
Submit Competitive Papers to: Top Paper Award
Receive a $500 cash award and the opportunity to be reviewed for publication in Journal of Personal Selling and Sales Management
Dr. David Shepherd, NCSM Program Chair Department of Marketing & Professional Sales College of Business Kennesaw State University 1000 Chastain Rd. #0406 Kennesaw, GA 30144-5591 (770) 423-6405 (770) 499-3261 Fax dshepherd@kennesaw.edu
Emphasis will be placed on:

Teaching Pedagogy Ethics Training Performance Evaluation Educational Issues Sales Motivation Selection of Sales Personnel Technology (e.g. CRM) Role of Communications in Sales
Submit Doctoral Student Papers to:

Dr. Mary Shoemaker, Chair NSCM Doctoral Student Sales Research Program Widener University One University Place Chester, PA 19013 (610) 499-4331
Hotel Information:
Marriott City Center 30 South Seventh St. Minneapolis, MN 55402 (612)-349-4000 800-932-2198
Sponsored by:
Pi Sigma Epsilon, Mu Kappa Tau and the University Sales Center Alliance
Note: When booking your room, remember to ask for the special Pi Sigma Epsilon convention rate.
Visit the NCSM Website at www.ncsmweb.org
Submission Deadline: September 30, 2005

34
35
The Communication Effectiveness of System Models Using the UML versus Structured Techniques:
A Field Experiment
Bruce C. Hungerford, University of Wisconsin-Oshkosh Michael A. Eierman, University of Wisconsin-Oshkosh
Abstract
The Unied Modeling Language has become an alternative to traditional modeling languages such as data ow diagrams for use in systems analysis. A modeling language is used to represent an information system so that analysts can use the model to make decisions about the design of the system and to communicate with stakeholders about the system. This study examines the comparative effectiveness of the UML and traditional modeling languages in communicating information about a system design. The study examines this on three types of individuals: individuals with no knowledge of either modeling language, individuals with no knowledge of either language that were provided training in one of the languages, and individuals that have had more extensive training in one of the languages. The study nds that there is no difference in the ability to communicate system design information between the languages for the rst two types of individuals. However, the study nds that, for more extensively trained individuals, systems modeled with the UML are better able to communicate information about the data in the system while systems modeled with traditional languages are better able to communicate information about the process used by the system.
to generate usable work products; (5) to manage information about large systems; (6) to explore multiple solutions economically; and (7) to master complex systems (Rumbaugh, Jacobson, and Booch 1999).
...a model is a vehicle for communicating with designers and users about the requirements and design of a proposed information system.
Introduction
The development of information systems for business is a costly, complex process that often does not lead to the desired system. To manage this development, organizations use formal methods and tools during the analysis of the potential system. Systems analysis involves gathering information to document the requirements of the system and capturing the essence of these requirements in a model. This model is then used, rened, and modied to develop a design for the proposed information system. The model is thus, the centerpiece of information systems development activity. The purposes of the model are varied and include: (1) to capture and state requirements and knowledge so that all stakeholders understand and agree with them; (2) to facilitate thinking about the design of a system; (3) to capture design decisions separate from the requirements; (4)
The Unied Modeling Language (UML) has received a signicant amount of attention as the tool of the future for modeling information systems. However, IS modeling was done prior to the development of the UML with tools such as Data Flow Diagrams (DFDs) and Entity-Relationship Diagrams (ERDs). These tools have been developed over many years of information systems development and are (and have been) taught to thousands of information systems professionals. Many organizations still use these tools, and many other organizations use these tools but are considering switching to the UML. While the UML is being promoted as the future of IS modeling, there is little empirical evidence that suggests it is better at fullling the purposes of a model as identied above. This research seeks to help decision makers make informed decisions about the adoption of UML over traditional modeling languages by examining the efcacy of the two primary approaches to information systems modeling with respect to their ability to fulll the rst purpose of a model. That purpose is to capture and state requirements and knowledge so that all stakeholders understand and agree upon them. In other words, a model is a vehicle for communicating with designers and users about the requirements and design of a proposed information system. The research specically compares models developed using the UML to those developed with traditional languages to determine their respective ability to communicate information about a proposed information system to users.
35
Hungerford and Eierman
Theory
Fowler and Scott (2000) suggest that the fundamental reason to use the UML is communication. Agarwal and Sinha (2003) suggest that UML diagrams enhance communication with stakeholders in a systems development project. However, little empirical research has compared UML to traditional modeling languages in terms of their relative effectiveness at communication. The limited research has focused on object-oriented development methods versus traditional analysis and design approaches and does not specically include the UML. However, based on this work, there is evidence that the type of modeling language may have an impact. In a study of the cognitive t between task and tool, Agarwal, Sinha, and Tanniru (1996) nd that process-oriented tools perform better than object-oriented tools on process-oriented tasks; however, object-oriented tools do not perform better than process-oriented tools on object-oriented tasks. In a later study, Krovi and Chandra (1998) nd evidence that an object-oriented model is easier to understand than a process model because it more closely resembles the cognitive representations used by individuals. Agarwal, De, and Sinha (1999) provide contrasting ndings in a study of comprehension of object-oriented and processoriented models. They nd that with easier questions about the model, there was no difference in the models. However, with more complex questions, the process-oriented model was more effective. These three studies used cognitive t as a theoretical basis for their investigation. Two of the studies examined the cognitive t between the model and the way humans think, while one examined the cognitive t between the model and the task. This research takes a different approach and examines the impact of object-oriented and process models on communication, utilizing research on message effects and graphics comprehension in communication theory to develop a theoretical basis for the study. Sager (1994) states that communication is a purposeful human activity concerned with effecting the knowledge structure of individuals. The message is the vehicle through which communication occurs (Bowers 1989). A message must be processed cognitively for the message to have an effect (Kellermann and Lim 1989). Therefore, if a model of an information system is used primarily to communicate with users of the information system, the model may be conceptualized as a message that must be processed cognitively by users to develop their understanding of the system. Furthermore, an information system model using traditional techniques such as DFDs and ERDs or the UML is a graphical representation of the system, which means the message for communication is diagrammatic. Larkin and Simon (1987) suggest diagrams may be superior to verbal descriptions because: Diagrams can group together all information that is used, thus avoiding large amounts of search activity for the elements needed to make an inference. Diagrams typically use location to group information
about a single element, avoiding the need to match symbolic labels. Diagrams automatically support a large number of perceptual inferences, which are extremely easy for humans. Winn (1994) presents an overview of how the symbol system of graphics interacts with the viewers perceptual and cognitive processes. In his description, the graphical symbol system consists of two elements: (1) Symbols that bear an unambiguous one-to-one relationship to objects in the domain of interest; and (2) The spatial relations of the symbols to each other. Thus, how symbols are congured spatially will affect the way viewers understand how the associated objects are related and interact. Zhang (1997), in an experiment using a Tic-Tac-Toe board and its logical isomorphs, shows that external representations of information are more than just memory aids. Her research suggests that the form of representation determines the information that can be perceived in a diagram. These studies suggest that different methods of graphically representing an information system may impact the ability of the user to comprehend the features and functioning of the system being modeled. Therefore, it is reasonable to assume that the difference in graphical representation of the UML and traditional modeling tools may have an impact on the effectiveness of communication with these tools. This research explores this idea in an attempt to understand if there is a signicant difference between the traditional modeling methods and the UML in effectiveness in communicating with users.
Hypotheses
The message in communicating with users about an information system is the model. A model is a type of diagram that must be processed cognitively to be understood. To process a diagram, the individual must use three types of processes (Larkin and Simon 1987): (1) Search, in which the user of the diagram seeks to locate specic sets of elements; (2) Recognition, in which the user matches elements located in the search with data required; and (3) Inference, in which the user adds new information to his or her understanding. This research specically focuses on investigating an individuals relative ability to recognize information in information system models developed with the UML and those developed with traditional modeling languages. For a user to comment effectively on the accuracy of the information system model, he or she must be able to identify and understand what is being communicated about the designers conception of the information system in specic areas. For example, if the user is asked if the data required about employees is complete, he or she has to be able to nd the area in the model that models information about employees and to identify the specic model component that holds this information. This process must take place prior to the users being able to make inferences about that information. Effec-
36
tive communication requires the receiver of a message to be able to recognize relevant information in the message. Although searching for that information is an important rst step in communicating using a diagram, this research does not examine differences in search between the approaches for two reasons. First, the ability to nd required information effectively will necessarily have a direct impact on differences in the other two processes. If individuals cannot easily nd the information they are looking for, they will not be as effective at either recognizing required information or at making inferences about that information. Second, effective investigation of search activity requires sophisticated technology and research techniques that we are not prepared to perform at this time. This focus means that if differences in the two models are identied, we will not be able to determine if the difference was due to the impact of the modeling languages on the search process or on the recognition process. However, any identied differences will imply a difference in the ability of the modeling languages to communicate information. This research also does not investigate the third process, Inference, for two reasons. First, the ability to make inferences is dependent on successful completion of the rst two processes. Second, the ability to make inferences about the information acquired from a model requires individuals who have sufcient knowledge about a situation to allow them to apply the new information. A portion of our subject population for this study does not have that knowledge. Accordingly, the current research examines if there is a difference in the ability of individuals using different types of information systems models to recognize required data. The basic hypothesis investigated in this study is: H1: There is no difference in the ability to recognize required information between individuals using traditional information system modeling tools (e.g., DFD and ERDs) and those using the UML. Agarwal, Sinha, and Tanniru (1996) found that process-oriented tools perform better than object-oriented tools on process-oriented tasks. Agarwal, De, and Sinha (1999) found that neither type of model performed better on structural questions. These competing ndings and a lack of more research in the area suggest that there is no good evidence that either type of model is better for modeling data or process in an information system. It is possible that traditional modeling, with its primary tool being the data ow diagram, would be better at communicating process information. The UML, with its primary tool being the class diagram, may be better at communicating information about data in the information system. When choosing a modeling language, it may be benecial to organizations to match the language to the dominant feature of the system to be developed to achieve better results. Systems where data is more important than process may be best developed with one language and systems where process is more important
than data may be best developed with another language. Therefore, this study examines two additional hypotheses to attempt to determine if either the UML or traditional modeling languages are better at communicating one type of information or the other. H2a: There is no difference in the ability to recognize required information about the data used in the information system between individuals using traditional information system modeling tools (e.g., DFD and ERDs) and those using the UML. H2b: There is no difference in the ability to recognize required information about the processes used in the information system between individuals using traditional information system modeling tools (e.g., DFD and ERDs) and those using the UML.
Method
The hypotheses were investigated with a series of three eld experiments using undergraduate students. Each experiment represented one of three different levels of training (none, limited, and extensive). Two treatments were given to each of these levels of training. The UML treatment consisted of a model of an information system that included a Use Case Diagram, a Class Diagram, a set of ve Sequence Diagrams, and a State Diagram. The traditional treatment consisted of a model of the same information system that included a Context Diagram, a level zero Data Flow Diagram, a set of four level one Data Flow Diagrams, an EntityRelationship Diagram, a Data Dictionary, and a Decision Tree. Both treatments included a questionnaire with twenty questions about the information system in the model, ve questions about the subjects perception of the task, and subject demographic questions. Since the focus of the study is on the users ability to recognize information in the model in the different types of models, the questionnaire had the same twenty questions for each treatment. The questions asked the user about different aspects of the information system, such as what data were needed and how different processes worked. Questions were multiple-choice. The treatment information system described by the models was an on-line grocery store. Both the UML and traditional model for this system were developed by the authors. Each model was reviewed for accuracy by an MIS professor not involved in the study. The instruments used in the experiment were then tested in three different pilots. The subjects in the pilots included students in introductory information systems classes at both the undergraduate and graduate level. The pilots focused on ensuring that both the UML model and the traditional model provided the same information and that each model provided enough information to answer all the questions on the survey. The pilots were
37
also used to test experimental procedures. Modications to each treatment model and the procedure were made after the rst two pilots.1 Experiment #1 No Training The rst2 experiment used students enrolled in Introduction to Marketing sections. These students did not receive any training on either modeling method. Subjects that had been exposed previously to one of the modeling methods were removed from the analysis. Subjects were assigned randomly to one of the two treatments described above: (1) system description with traditional modeling diagrams, or (2) system description with the UML. The experimental session lasted forty minutes. All students received extra credit in their marketing class for participation in the experiment. Students were informed that subjects who scored above 70 percent on the survey would be put in a pool from which four individuals (two from each treatment group) would be chosen at random to receive a $20.00 prize. The experiment was administered in ve sessions. Both modeling languages were tested in the same session by randomly assigning the subjects to use one or the other treatment. Of the sixty-two students that participated in the experiment, four were eliminated during analysis because they indicated that they had been exposed to the modeling language prior to the experiment. Experiment #2 Minimal Training The second experiment also used students enrolled in Introduction to Marketing sections. However, for this experiment, the subjects received limited training on the modeling method to which they were assigned randomly. The experiment lasted approximately one hour. In the rst twenty minutes of the experimental session, the subjects read a training document that included an explanation of the symbols used in the modeling language treatment to which they were assigned, a written description of an information system, and a model of an information system that was an-
notated to explain how the model represented the information system. The nal forty minutes of the session followed the same protocol as the rst experiment. Students were rewarded with the same approach as in the rst experiment. The training document used in the second experiment was developed to explain the symbols used in each of the modeling languages and to provide an example of their use. For the traditional modeling language, the training document included Data Flow Diagrams, an Entity-Relationship Diagram, a Decision Tree Diagram and a Data Dictionary. For the UML, the training document included Use Case Diagrams, Class Diagrams, Sequence Diagrams, and State Diagrams. The documents were developed to provide the same level of explanation for each modeling language, and were reviewed for accuracy and completeness by two MIS professors not involved in the project. The training information system was a Student Registration System. This type of system was selected because it was believed that students would have some knowledge of how such a system worked and would therefore nd it easier to understand the example. In the rst pilot, the training documents were provided to the subjects, but were explained by one of the authors. It was determined from this experience that we could not guarantee the same level of training between experimental sessions because of student questions and level of detail covered. We decided, therefore, to provide self-paced training by allowing subjects to examine the training documents on their own. No instruction on either of the modeling languages was provided by anyone during the experimental sessions. Students studied the training document on their own during the rst twenty minutes of the session. The experiment was administered in four sessions. Both modeling languages were tested in the same session by assigning the subjects randomly to use one or the other treatment. Of the sixty-eight students that participated in the experiment, ten were eliminated during analysis because they indicated that they had been exposed to the modeling language prior to the experiment.
Table 1 Summary of Treatments
Experiment 1
Subjects Marketing Students (n=58) Marketing Students (n=58) MIS Students (n=101) None
Training
Treatment System Model of On-Line Grocery Store System Model of On-Line Grocery Store System Model of On-Line Grocery Store
Description Five 40-minute sessions with subjects randomly assigned UML or Structured Treatment Four 60-minute sessions with subjects randomly assigned UML or Structured Treatment
System Model of Student Registration System. Training Document Explaining Artifacts Course Work
Students assigned treatment based on the course in which they were enrolled. Each treatment was part of their nal exam.
38
Experiment #3 Extensive Training The third experiment examined the impact of more extensive training on the communication effectiveness of the modeling language. This experiment used students enrolled in a systems analysis class and students enrolled in a UML and Java class. All students in these classes were MIS majors. Students were not enrolled in both classes. All students in the UML/Java class had previously taken the systems analysis class. Training on the tools in each of the classes was approximately equal in terms of time spent and work performed. In the systems analysis class, students received instruction on the use and meaning of traditional development tools including Data Flow Diagrams and Entity Relationship Diagrams, and then practiced using these modeling tools to describe systems. In the UML and Java class, students received instruction on the use and meaning of UML diagrams and then practiced using these tools to describe systems. Both classes included graded assignments where students developed system models using the respective tools. Students in each class completed the treatment as a part of the nal exam for the course. In this case, treatment was also the on-line grocery store. However, the specic treatment was assigned based on the class in which the student was enrolled rather than randomly. Data was collected from 101 subjects. Sixty-two subjects in two classes received the traditional language treatment and thirty-nine subjects in one class received the UML treatment. Since subject recruitment was dependent on class enrollment it was not possible to control sample size. However, the unequal sample size should not make a statistical difference because both treatments had more than thirty observations and the only statistical tests performed examined differences in the means of the two treatments (Sincich 2004; McClare and Sincich 2002). After the data were collected, preliminary analysis showed that one question was answered incorrectly by all students with extensive training and no training that received the UML treatment. A large number of students that received the UML treatment and minimal training also missed the question. Subsequent investigation showed that the information required by the question was not included in the model. The question was eliminated from the analysis for both treatments at all training levels. The three experiments are summarized in Table 1.
subject with a score of 35 percent. Statistical tests were performed on the subjects overall scores, scores on questions concerning the data represented in the model (a total of 9 after adjustment), and scores on questions concerning the process represented in the model (10). The normality of the output for overall, data, and process scores was checked, and all met the skewness and kurtosis requirements. Each hypothesis was then examined using the appropriate t-test (Morgan and Griego 1998). The results of a Mann-Whitney U test also are reported because the data for subjects with extensive training did not meet normality requirements. Therefore, use of a non-parametric test was appropriate for that analysis. Results for these subjects are summarized in Table 2.
Table 2 ResultsSubjects with No Training
UML Traditional t-test (p) MannWhitney U (p) 353.5 (0.3060) 422 (0.9551) 353.5 (0.3024)
Subjects Average Score All Questions Average Score Data Questions Average Score Process Questions
31 0.573 0.606 0.543
27 0.555 0.606 0.507 0.52 (0.6019) -0.01 (0.9936) 0.73 (0.4660)
Results
Subjects with No Training Data analysis was performed on fty-eight questionnaires: twenty-seven subjects received the traditional modeling language treatment, and thirty-one subjects received the UML treatment. The subjects in each treatment were demographically similar in all aspects. The hypotheses were tested using a single factor, two treatment level design. The dependent variables are based on the scores of the subjects. They are ratio in nature a subject with a score of 70 percent did twice as well as a
The average score for subjects on all questions was 57.3 percent for the UML treatment and 55.5 percent for the Traditional treatment. Test statistics demonstrate that there is not a signicant difference between the mean scores. Therefore, the rst hypothesis cannot be rejected for subjects with no training. The average score for subjects on data questions was 60.6 percent for the UML treatment and 60.6 percent for the Traditional treatment. Test statistics demonstrate that there is not a signicant difference between the mean scores. Therefore, the second hypothesis cannot be rejected for subjects with no training. The average score for subjects on process questions was 54.3 percent for the UML treatment and 50.7 percent for the Traditional treatment. Test statistics demonstrate that there is not a signicant difference between the mean scores. Therefore, the third hypothesis cannot be rejected for subjects with no training. Subjects with Minimal Training Data analysis was performed on fty-eight questionnaires: thirty subjects received the traditional modeling language treatment, and twenty-eight subjects received the UML treatment. The subjects were demographically similar in all aspects except gender. A disproportionate number of females received the UML treatment. Gender differences
39
Table 3 ResultsSubjects with Minimal Training
Table 4 ResultsSubjects with Extensive Training
UML
Traditional
t-test (p)
MannWhitney U (p) Subjects 62 368 (0.4153) 403 (0.7867) 367 (0.4045) Average Score All Questions Average Score Data Questions Average Score Process Questions
UML
Traditional
MannWhitney U (p) 1294.5 (0.5467) 757 (0.0011) 1701 (0.0005)
Subjects Average Score All Questions Average Score Data Questions Average Score Process Questions
28 0.609 0.661 0.562
30 0.571 0.630 0.517 1.08 (0.2843) 0.71 (0.4829) 0.92 (0.3640)
39 0.772 0.764 0.823
0.778 0.876 0.725
were tested, and no signicant difference in performance between male and female subjects was found. The hypotheses were tested in the same manner as used in the subjects with no training. Normality of the output also met requirements for this data set. Variance assumptions were checked for each hypothesis using Levenes Test for Equality of Variances; variance equality could not be rejected at the .05 level for overall and data scores, but was rejected for the process scores. The appropriate t-test was used. Results for these subjects are summarized in Table 3. The average score for subjects on all questions was 60.9 percent for the UML treatment and 57.1 percent for the Traditional treatment. Test statistics demonstrate that there is not a signicant difference between the mean scores. Therefore, the rst hypothesis cannot be rejected for subjects with minimal training. The average score for subjects on data questions was 66.1 percent for the UML treatment and 63.0 percent for the Traditional treatment. Test statistics demonstrate that there is not a signicant difference between the mean scores. Therefore, the second hypothesis cannot be rejected for subjects with minimal training. The average score for subjects on process questions was 56.2 percent for the UML treatment and 51.7 percent for the Traditional treatment. Test statistics demonstrate that there is not a signicant difference between the mean scores. Therefore, the third hypothesis cannot be rejected for subjects with no training. Subjects with Extensive Training Data analysis was performed on 101 questionnaires: thirty-nine subjects received the traditional modeling language treatment, and sixty-two subjects received the UML treatment. Gender differences were tested, and no signicant difference in performance between male and female subjects was found. The hypotheses were tested in the same manner as used in the subjects with no training. However, normality of the output did not meet all the skewness and kurtosis requirements. Therefore, tests for differences were performed with 40
the Mann-Whitney U test. Results for these subjects are summarized in Table 4. The average score for subjects on all questions was 77.8 percent for the UML treatment and 77.2 percent for the Traditional treatment. Test statistics demonstrate that there is not a signicant difference between the mean scores. Therefore, the rst hypothesis cannot be rejected for subjects with extensive training: There is no difference in the ability to recognize required information between individuals using traditional information system modeling tools (e.g., DFDs and ERDs) and those using the UML. The average score for subjects on data questions was 87.6 percent for the UML treatment and 76.4 percent for the Traditional treatment. Test statistics demonstrate that there is a signicant difference between the mean scores at the 0.0011 level. Therefore, the second hypothesis can be rejected for subjects with extensive training: There is a difference in the ability to recognize required information about the data used in the information system between individuals using traditional information system modeling tools (e.g., DFDs and ERDs) and those using the UML. The average score for subjects on process questions was 72.5 percent for the UML treatment and 82.3 percent for the Traditional treatment. Test statistics demonstrate that there is a signicant difference between the mean scores at the 0.0005 level. Therefore, the third hypothesis can be rejected for subjects with extensive training: There is a difference in the ability to recognize required information about the process used in the information system between individuals using traditional information system modeling tools (e.g., DFD and ERDs) and those using the UML. Training The objective of this study was to examine the difference in communication effectiveness of the two different types of modeling languages. Both end-users and developers use system diagrams during the analysis and design of an information system. When users are involved in a systems project they may be asked to comment on the analysis or design of the system as described by the system model. The
level of training needed for users to effectively comment on the model is an issue of potential importance. Therefore, the level of training received by subjects in the experiment also served as an independent variable. The impact of training on communication effectiveness of the models was tested with a two-way ANOVA. The overall average score for subjects with no training, minimal training, and signicant training were 56 percent, 59 percent, and 77 percent respectively. The overall average score for all subjects receiving the UML treatment was 65 percent while the overall average score for all subjects that received the traditional treatment was 63 percent. Test statistics demonstrate that training had a signicant impact while language type did not. The results were no different when the data and process scores were examined independently of one another. The related statistics are provided in Table 5.
Table 5 Two-way ANOVA Results
N Training No Training Minimal Training Signicant Training 2 2 2 0.5641 0.5901 0.7749 0.0122 0.0271 0.0045 0.00864 0.01914 0.00318 Mean SD SE
In both languages, subjects without extensive training found it somewhat difcult to relate the diagrams to develop an overall understanding of the system and somewhat difcult to nd specic information to answer the questions. The questions and answers for both the UML and traditional languages at all training levels received an almost identical evaluation of being somewhat easy to understand. The questions and answers were exactly the same for both languages, which suggests that subjects understood what they were supposed to do and implies that other differences detected in the study are attributable to differences in the modeling languages.
Discussion
This study nds that training had a signicant impact on the ability of users to identify information in the models of both modeling languages examined. Those with extensive training were better able to identify required information than those individuals with little or no training. The study also nds a signicant difference between the modeling languages for extensively trained individuals in the type of information that was required to be identied. Extensively trained individuals using the Unied Modeling Language were better able to identify information about the data associated with an information system than were users of the traditional modeling languages. In contrast, extensively trained individuals using the traditional modeling languages were better able to identify information about the processes associated with an information system than were individuals using the Unied Modeling Language. It is difcult to determine why these differences occurred. Given that the traditional model included both an Entity-Relationship Diagram and a data dictionary, whereas the UML model included only a class diagram, one might expect that subjects using the traditional model would perform better on data questions because the information was located in more than one place in the model. Even a random search for answers to data questions would more likely lead to a document that could provide an answer in the traditional model because data was represented in more places. This suggests there could be an inherent advantage to representing data in a class organization over presenting it in an entity-relationship organization. As Zhang (1997) suggests, the form of the diagram impacts the ability to perceive information, and therefore the form of the class diagram allows better perception of information than the form of the ER diagram. This nding contradicts the ndings of Agarwal, De, and Sinha (1999), who found that neither model type performed better on structural questions. A possible explanation for these results may be found in the task-technology t (TTF) literature (e.g., Goodhue 1995; Goodhue and Thompson 1995). This literature suggests that software tools will be used more if they meet the needs of the task because the user perceives a better ability to complete the task with the tool than without it. While the lit41
Language UML Traditional 3 3 0.6533 0.6327 0.1095 0.1206 0.06324 0.06965
Source of Variation Training Language Within Cells Total
SSq 0.0528 0.0006 0.0003 0.0537
DF 2 1 2 5
MSq 0.0264 0.0006 0.0001
F 200.64 4.85
P 0.00496 0.15847
Perception of the Documents and Task Data was collected about subjects perception of the task and the diagrams used in the models. The subjects perceptions are examined to determine if the subjects perceived one treatment to be more difcult than the other. A difference identied in perceived difculty could indicate aws in either the treatment documents or the training documents that would skew the validity of the results reported above. The analysis determined that there were no signicant differences between the traditional modeling language and the UML in subjects perceptions of the task, the questions, or the diagrams. However, the extensively trained individuals appeared to nd the task easier than did the other subjects.
erature does not test actual performance, it does suggest that better performance will result when the technology ts the task. Dishaw and Strong (2003) extend the TTF literature by examining the role experience with the task and experience with the tool have on the perception of task-technology t and tool utilization. Their study found that experience with the tool leads to a better t and higher tool utilization. They did not nd a relationship between experience with the task and t or utilization. While this paper does not explicitly study the t between the information communication task and modeling language, the theory may be used to explain the results. For untrained and minimally trained individuals there was no task-technology t. They did not have enough experience to effectively use either tool in an information identication task. However, more extensively trained individuals could have found a t between the task and the technology for different aspects of the task because they performed better with one tool in one aspect of the task than they did with the other tool. This suggests traditional modeling languages may have a better t with tasks concerning identication of process information in an information system model, while the UML may have a better t with tasks concerning the identication of data information in an information systems model.
use in information system design and the impact of modeling language on these uses.
Notes
1. 2. The questionnaire and treatment materials are available from the authors. Refers to logical order; this experiment actually was conducted after the second experiment due to subject availability, etc.
References
Agarwal, R., P. De, and A.P. Sinha. 1999. Comprehending object and process models: An empirical study. IEEE Transactions on Software Engineering 25(4):541-555. Agarwal, R. and A.P. Sinha. 2003. Object-oriented modeling with UML: A study of developers perceptions. Communications of the ACM 46(9):248-256. Agarwal, R., A.P. Sinha and M. Tanniru. 1996. Cognitive t in requirements modeling: A study of object and process methodologies. Journal of Management Information Systems 13(2):137-162. Bowers, J.W. 1989. Introduction in Message Effects in Communication Science. 10-23. Ed. J.J. Bradac. Newbury Park: Sage Publications. Dishaw, M.T.and D.M. Strong. 2003. The effect of task and tool experience on maintenance CASE tool usage. Information Resources Management Journal 16(3):1-16. Fowler, M. and K. Scott. 2000. UML distilled: A brief guide to the standard object modeling language. New York: AddisonWesley. Goodhue, D.L. 1995. Understanding user evaluations of information systems. Management Science 41(12):1827-1844. Goodhue, D.L. and T.L. Thompson. 1995. Task-technology t and individual performance. MIS Quarterly 19(2):213-236. Kellermann K. and T. Lim. 1989. Inference-generating knowledge structures in message processing. Message Effects in Communication Science, 102-128. Ed. J.J. Bradac. Newbury Park: Sage Publications. Krovi, R. and A. Chandra. 1998. User cognitive representations: The case for an object oriented model. The Journal of Systems and Software 43:165-176. Larkin, J.H. and H.A. Simon. 1987. Why a diagram is (sometimes) worth ten thousand words. Cognitive Science 11:65-99. McClara, J. and T. Sincich. 2002. Statistics. 9th ed. Pearson Education. Morgan, G.A. and O.V. Griego. 1998. Easy use and interpretation of SPSS for windows: Answering research questions with statistics. Mahwah: Lawrence Erlbaum Associates. Rumbaugh, J., I. Jacobson and G. Booch. 1999. The unied modeling language reference manual. N.Y.: Addison-Wesley. Sager, J. 1994. Language engineering and translation: Consequences of automation. Amsterdam: John Benjamins. Sincich, T. Personal communication with author. July 8, 2004.
Conclusion
The study has obvious limitations. First, the subjects were undergraduate students with no signicant prior knowledge of the information system presented to them. This may not be a realistic representation of users involved in the design of a system. Second, the subjects were not able to ask questions or discuss the design with an individual knowledgeable about the design. This also is probably not representative of the real world. Finally, the system used in the research, while not trivial, was not of the size and complexity found in systems typically developed by business. Despite the limitations, this study does imply that moving to the UML as a modeling language for information systems may not confer a tremendous advantage over using traditional modeling languages in communicating information about the system. However, if the users are given a signicant amount of training, the UML model may be more effective in modeling systems with a strong data focus, while traditional models may be more effective in modeling systems that are strongly process oriented. The implication for this study is that if a business is building an information system where process is the primary concern of the designers, then, at least for communication purposes, it should use a traditional modeling language. On the other hand, if the important aspects of the systems design has a strong data focus, then using the UML may be a better choice. More studies are needed that explore different aspects of model
42
Winn, W. 1994. Contributions of perceptual and cognitive processes to the comprehension of graphics. Comprehension of Graphics, Ed. W. Schnotz and R.W. Kulhavy. Amsterdam: North-Holland. Zhang, J. 1997. The nature of external representations in problem solving. Cognitive Science 21(2):179-217.
About the Authors

Bruce C. Hungerford received a Ph.D. degree in information systems from the University of South Florida. He is an Assistant Professor with the Management Information Systems Team in the College of Business Administration at that University of Wisconsin-Oshkosh. His research interests include information systems development, software engineering, software quality improvement, and enterprise resource planning systems. Prior to earning his doctorate, he worked in the nancial and IT industries. He is a member of the ACM, AIS, and IEEE Computer Society. hungerfo@uwash.edu Michael Eierman is an Associate Professor of the Management Information Systems Team in the College of Business Administration a the University of Wisconsin-Oshkosh. He earned a Ph.D. in Management Information Systems from the University of Minnesota. His current research focus is on object-oriented technologies including the impact of UML on systems analysis and maintenance of object-oriented software. eierman@uwash.edu
43
Call for Manuscripts

Information Technology, Learning, and Performance Journal
OSRA: THE ORGANIZATION
The Organizational Systems Research Association brings together professionals from the business and academic worlds with a focus on information technologies and their impact on learning and performance. OSRA is small enough so that you can really get to know colleagues and feel a part of things at the conference. But its also big enough to attract distinguished papers and speakers.
KEY ACTIVITIES
Research Conference
An annual event attracting experts in the eld of organizational implementation of information technology, business education, corporate training, and curriculum development. Refereed conference papers are presented and published in the online conference proceedings.
Article Submission:
Submit 5 copies of your manuscript to: Dr. Susan Feather-Gannon Pace University-CSIS 826 Van Wagner Road Poughkeppsie, NY 12601-6504 To obtain a sample issue of the Journal, contact: Donna R. Everett (606)783-2718 d.everett@moreheadstate.edu
www.osra.org
A website of OSRAs publications in electronic forms, plus updates on OSRA activities, jobs, grants, and research in progress.
Publishing Opportunity
InformationTechnology,Learning,and Performance Journal A biannual publication (included with membership) presenting refereed research articles in the eld of organizational and end-user information systems. The Making a Difference section explores policies, methods, and best practices relevant to technology professionals, educators, and trainers.
OSRA Newsletter
A quarterly publication (included with membership) to update members on OSRA activities and events, as well as jobs, grants and research in progress.
44
Building a Market-Oriented Organizational Environment:

An Implementation Framework for Small Organizations
Beth Ann Martin, John Carroll University James H. Martin, John Carroll University
Abstract
The strong link between a market orientation and performance in small organizations rests on the organizations ability to use its market-oriented culture to create a sustainable competitive advantage. To do this requires the rm to build and maintain a strong market orientation. Using an internal customer-internal supplier perspective, this paper identies a framework for implementation that an organization can undertake to create a market-oriented workforce. The foundation for the framework is the development of dyadic relationships between internal customers and suppliers. The implementation structure relies on a performance management system that rewards behaviors appropriate for the establishment of a market-oriented culture.
cess by which an organization would implement a market orientation. Although there is some guidance in terms of general areas of concern (e.g., Jaworski and Kohli 1993; Harris 1999) very few published studies articulate a systematic approach that ties the necessary concepts together into an overall framework. Based on a model of Management Intervention suggested by Porras and Robertson (1990) we propose a series of implementation activities that are grounded in theory, testable and managerially focused.
...smaller organizations frequently must rely upon limited resources to be competitive, suggesting a market-oriented culture can be an exceptionally important resource for the small organization.
Introduction
A resource-based view of the organization (Chatterjee and Wernerfelt 1991; Hunt and Morgan 1995; Wernerfelt 1989) suggests organizations should focus their efforts on developing and maintaining resources that will help the organization develop strategic skills and capabilities for implementing value-creating strategies. Pelham and Wilson (1996) and others (e.g., Day 1994; Slater and Narver 1999) argue that a market-oriented organizational culture can be an especially strong resource for developing strategies that lead to increased performance. Pelham and Wilson (1996) studied smaller rms and found that the inuence of the organizations strategy and structure had less impact on performance than did having a market-oriented culture. Thus, while larger organizations have a fairly wide and substantial base of resources from which to draw (e.g., nancial, human, technological, etc.), smaller organizations frequently must rely upon limited resources to be competitive, suggesting a market-oriented culture can be an exceptionally important resource for the small organization. The issue confronting an organization wanting to develop a market-oriented culture is how to bring this about. While the nature of the relationship between market orientation and rm protability has received substantial attention in the literature, there has been less attention given to the pro-
A systematic approach to improving organizational performance through the development of a market oriented culture can be considered a management intervention. Management interventions can be viewed as planned changes in a work setting that are designed to change the behavior of individual organization members and ultimately lead to improved organizational outcomes (Porras and Silvers 1991). A model of the management intervention process described by Porras and Robertson, (1990: Figure 1) presents organizational intervention as potentially producing changes in a variety of workplace characteristics. The basic assumption in the model is that changing the work setting (including social factors, technology and physical characteristics) is the most inuential tool for changing individual behavior, which in turn should lead to improved organizational outcomes. The foundation for the model rests on the perspective that altering the work setting can introduce changes in individual behavior. This assumption is based on cognitive models of behavior which postulate that an individuals environment is an important source of information about appropriate behavior (Porter and Lawler 1968; Hackman 1981). It is proposed that a management intervention that focuses on the organizational factors inherent in a market oriented organization will result in changing individual worker behavior to become more market oriented. 45
Martin and Martin
Developing a market orientation is a management intervention. As our discussion develops, it will become clear that there are necessary changes in the organizational setting, social factors and possibly technology that will lead to changes in employee attitudes and behavior toward becoming a market oriented organization. We will begin by providing an overview of market orientation. While this is not a comprehensive literature review, it will highlight some of the main issues and concerns in this literature. Subsequent sections will discuss using an organizations performance management system and an internal customer orientation as the basic mechanisms for the framework. The nal section will provide a 6-step model for initiating a market orientation.
Literature Review
Market Orientation and Performance Over the years there has been conict and debate in the conceptualization of market orientation. One view of market orientation has presented the concept from an information/attitudinal perspective (Han, Namwoon and Srivastava 1998; Hooley, Lunch and Shepherd 1990; Narver and Slater 1990). This perspective suggests that market oriented organizations are interested in, actively maintain, and use information about customers, competitors and general market trends. The original focus for this perspective was the type of market information the company maintained and the degree to which employees were interested in it. A second approach to market orientation has been a behaviorally focused conceptualization (Deshpande 1999; Jaworski and Kohli 1993; Kohli and Jaworski 1990). This approach has suggested three behavioral elements are required for an organization to function with a market orientation. The rst, intelligence generation, includes customers verbalized needs and preferences as well as the analysis of exogenous factors that inuence customers needs and preferences. The second behavioral element of a market orientation is intelligence dissemination, which is the communication of the information throughout the organization. Providing all employees across the organization with market information is important because it facilitates the third element, responsiveness to market intelligence. Response to market information is considered to be more rapid and more effective when the organization as a whole is knowledgeable about the needs of the competitive marketplace. In a discussion of the attitudinal/behavioral issues, Avlonitis and Gounaris (1997) have offered a reconciliation between these two conceptualizations by suggesting that an organization must emphasize both attitudes and behavior in its market orientation in order for the organization to be able to maintain its market orientation in the long run. Both Avlonitis and Gounaris and Wrenn (1997) conclude that there must be a combined focus on attitudes and practice (behavior). Based on Figure 1, the proposed framework strongly integrates these two perspectives by placing the initial emphasis on altering the organizational setting and, 46
thereby the employee attitudes, followed by an emphasis on individual behavior change. All sides to this debate agree that having a market orientation is not something an organization has or does not have. An organization is more or less market oriented, being able to respond better or less well to market opportunities. While the nature of market orientation and its role in improving business performance has been debated for over 30 years, only recently has it been empirically investigated. Several studies have found a positive relationship between market orientation and business performance (e.g., Pelham 2000; Deshpande and Farley 1999; Appiah-Adu and Ranchhod 1998; Appiah-Adu 1997). Narver and Slater (1990) and Slater and Narver (1994, 2000) report a positive relationship between market orientation and return on assets. In addition, Slater and Narver (1994) also suggested a positive relationship between market orientation and sales growth. Deshpande, Farley and Webster (1993) found businesses customer orientation is positively related to business performance. Appiah-Adu (1997), studying small rms, reported a positive impact of market orientation upon business performance, and Appiah-Adu and Ranchhod (1998) indicated market orientation is signicantly correlated with growth in market share, overall performance and prot margin, although not with new product success. Also focusing on small rms, Pelham and Wilson (1996) suggested a strong inuence of market orientation on measures of small-rm performance. For example, they reported market orientation positively inuenced the current years level of protability and they found market orientation was signicantly related to product quality, which was signicantly associated with growth share and protability. They concluded that a high level of market orientation can provide a small organization with a strong source of competitive advantage. Pelham (2000) reported that market orientation in small organizations is correlated positively with marketing/sales effectiveness, growth and protability. Recently, research has suggested that the impact of market orientation on an organizations performance may be moderated by such factors as the strategic direction of the organization (Matsuno and Mentzer 2000) or by extreme economic volatility (Grewal and Tansuhaj 2001). Mounting evidence at the organizational level has generally supported the relationship between market orientation and business performance, leading to a generally accepted conclusion that within certain constraints, an organization is better off with more market orientation than with less market orientation. Missing in the Literature While understanding the need for a market orientation is important, business leaders must also understand how to go about creating or developing a market orientation within their business. Although the marketing literature is beginning to show strong evidence of the positive impact of a market orientation on performance, only limited literature has investigated the issue of improving an organizations
Martin and Martin
use of market intelligence and this literature is lacking an integrated framework to help organizations establish a market orientation. Wrenn (1997) and Han, Namwoon and Srivastava (1998) suggest a paucity of research exists on how to manage and develop a rms market orientation. Early work in this area by Kohli and Jaworski (1990) suggests the critical need for upper management support and its impact on shaping organizational values central to a marketing culture. Garver and Cook (2001) discuss how companies can effectively use customer value and satisfaction data. They see the main challenge as incorporating customer satisfaction data to drive improvement. In contrast to previous market orientation frameworks, Garver and Cook suggest a customer value and satisfaction culture will focus on both attitudes and behaviors. It is this market-oriented culture that guides employees through the myriad of customer related data. They believe there are two equally vital processes; getting data to people and getting people to use the data. It is the response to the data that drives competitive advantage. The ideas proposed by Garver and Cook are clearly foundational building blocks for a model of developing a culture of market orientation. Based on the management intervention model above (Porras and Robertson 1990), along with the work of Garver and Cook (2001), this paper will identify tools and processes that can help an organization develop a market orientation in such a way that it becomes pervasive throughout the organization. The processes were designed so that small organizations with limited resources could pursue the development of a strong market orientation. While this model will provide direction for organizations of all sizes, the proposed steps can be most directly applied to small organizations. First, a dyadic perspective on an organization that includes an internal customer orientation will be discussed. Following this, a multi-level implementation framework will be developed and procedures will be outlined that can facilitate the creation of a market oriented organization. Developing An Internal Customer Orientation A market orientation is a shared set of values, beliefs and behaviors that focus on putting the customer rst (Deshpande 1999). The difculty with developing this orientation in practice is that employees are not likely to be willing to adopt these values, beliefs and behaviors without a specic structure and system to encourage change. A mechanism is needed that can be used for creating market oriented behaviors throughout the organization, along with the means for implementing the mechanism effectively. The proposed framework combines an internal customer orientation with a performance management system for that mechanism. The main focus from Porras and Robertsons model (1990), will be on changes to the organizational setting and individual behavior change. Baker, Simpson and Siguaw (1999) and Siguaw, Simpson and Baker (1998) report evidence that the relationship between external customers and external suppliers is stron-
ger when there is a match between the degree of market orientation of both customers and suppliers. Steinman, Deshpande and Farley (2000) report nding a gap between external customer and external supplier perceptions of each others market orientation, but that this perception gap lessens with lengthier and stronger relationships. Research generally shows that the stronger the relationship between customers and suppliers, the stronger the performance of both supplier and customer (Cannon and Homburg 2001). The total quality management literature has suggested that an organization can be thought of as interrelated sets of dyads between internal customers and internal suppliers (Goetsch and Davis 1997; Hallowell, Schlesinger and Zornitsky 1996). When internal suppliers are oriented toward satisfying their internal customers needs, organizational performance will be improved (Achrol and Kotler 1999). When the internal customer needs are dened in terms of the ability to help the organization satisfy external customer needs, then the result should be an organization that has become more market oriented. Hauser, Simester and Wernerfelt (1996) and others (e.g., Conduit and Mavondo 2000; Gronroos 1990) have suggested that in order to develop a market orientation, a rm must focus its internal suppliers on serving their internal customers, who in turn serve other internal customers who eventually serve external customers. Hauser, Simester and Wernerfelt conclude that having an internal customer focus is imperative to drive a market orientation deep within an organization. The key to bringing this about within their model is the organizations reward system. They propose a bonus type compensation system in which internal customers negotiate with their internal suppliers for payment based on the suppliers ability to help the internal customer satisfy their customer. Although the specic form of compensation may be less important as different organizations will want to work within different forms of compensation, the work of Hauser, Simester and Wernerfelt does pinpoint the necessity of tying compensation and other performance management activities to the specic, market oriented behaviors the organization is trying to elicit from its employees. Using an internal customer orientation to drive a market orientation deep within an organization requires the employees to learn how to focus on both the needs of the external customer and to perceive other employees as internal customers (Mohr-Jackson 1991). This necessitates a change in the way employees perceive their jobs. For example, within a typical company, employees may focus on the work they produce because the reward structure is based on productivity. An internal customer/supplier orientation suggests that, rather than focusing on the production of a certain number of components per hour, a component line producer would want to know if the person within his/her organization receiving the components was satised with the components quality and timeliness. However, the component maker is also a customer who needs to be satised. S/he is a customer of the department or person who delivers the materi47
Martin and Martin
als that are used to make the components. If the materials arrive late or with defects, then the component maker will be a dissatised internal customer. The component maker is part of at least two dyads. In the rst dyad the product component maker is an internal supplier of components to others, and in the second dyad s/he is an internal customer of other internal suppliers. Heilmann (1994) suggested that the internal customer focus is important because internal customer objectives can be aligned with rm objectives that would include providing superior customer value to external customers. This alignment of internal customer objectives with the objectives of the rm can take place through the implementation of a reward system and the implementation of a performance evaluation system to support the reward structure that includes the appraisal of the effectiveness of specic outcomes by internal customers (Conduit and Mavondo 2000; Hauser, Simester and Wernerfelt 1996). By working within the dyadic system it becomes apparent how a change in the organizational setting and social factors will impact individual behavior. Sustaining the Intervention: A Performance Management System A performance management system is one mechanism that can be used for creating and sustaining market oriented behaviors throughout the organization. There is ample evidence demonstrating that behavior-based performance evaluations are consistent with improved customer service (Anderson and Oliver 1987; George 1990). When organizational rewards are tied to specic employee behaviors a performance management system is created. Behavior based evaluations appraise employees on how they act rather than on the specic outcomes achieved. Behavior such as contacting members of ones own customer/supplier dyad to assess satisfaction, giving feedback to an immediate supplier regarding the timeliness or the quality of materials, and contacting ones own customers to assess his/her level of satisfaction would be behaviors critical to the success of developing a market oriented culture. As such, these behaviors would form the criteria upon which the actual performance appraisal is based and would be included on each employees performance evaluation. Identifying the specic behaviors involved in focusing the internal suppliers on serving their internal customers, rather than on objective criteria such as number of units produced, will create a clear customer oriented focus within the organization. Such a system gives employees the incentive to engage in behaviors that are supportive for improving product and service quality. Any desired change in the attitudes and behaviors of employees must be managed by altering the performance management system to reect the desired changes. While there are multiple ways to develop a performance management system, one alternative to achieve the balance between changing employee attitudes and subsequently changing employee behavior is to develop a process similar to the balanced scorecard proposed by Kaplan and Norton 48
(1992). The balanced scorecard is a management intervention system that provides a presentation of both nancial and operational measures to upper management so that they can receive a quick comprehensive view of the company. Because most activity takes place at department levels, all measures are decomposed to what are called local levels. Local level metrics allow top management to keep abreast of key internal processes that affect overall corporate objectives. A main benet of this process is that employees at even the lowest levels in an organization have clear targets and goals set for them, and these goals always contribute to the companys overall mission. In addition, weak performance is easy to diagnose. Setting behavior and/or performance targets for employees clearly communicates the attitudes and expectations of upper management. This is turn should lead to behavior change on the part of the workers such that their behaviors are in line with the goals of top management. Although a complete revamping of the rms performance management system may be an ideal, it is not necessary that a new performance management system be developed prior to initiating a market orientation program. A current performance management system can be amended to include metrics that assess employees gathering, dissemination and responsiveness to appropriate market intelligence. In addition, these measures would also need to contribute to each employees overall performance evaluation. Initiating a Market Orientation Based on the above discussion, moving an organization to a market orientation requires the achievement of two management objectives: a change in employee attitudes and a change in employee behavior. The rst objective is to change employee knowledge and attitudes in a way that reects a pervasive market oriented culture. However, changing knowledge and attitudes is not sufcient to build a long lasting market oriented culture. The second objective is to change employee behavior in a way that focuses the employee (as an internal supplier) on helping his/her internal customers better satisfy their customers in a way that ultimately helps the organization to satisfy external customers. To achieve these objectives, the organizations members must rst be made aware of the importance of internal customers and the organization must then identify the specic behaviors of employees that need to change. Reecting on the intervention model in Figure 1, emphasis will clearly be on the middle two steps of the model (changing the organizational setting and changing individual behavior). That emphasis will necessarily come from upper management intervention. The major recurring behavioral phases of a market orientation as described by Kohli and Jaworski (1990) and Jaworski and Kohli (1993) include gathering market intelligence, disseminating the market information and rapidly responding to the market information. Using these three phases as a guide, we proposed that new employee behaviors that must
Martin and Martin
Figure 1 Organizational Outcomes as a Function of Management Intervention
Management Intervention
Changes in: Organizing Arrangementsgoals, policies, reward systems Social Factorsmanagement style, informal networks, patterns Physical Setting Technology
Individual Behavior Change
Organizational Outcomes
Based on Porras and Robertson, 1990
be introduced into a rm with little or no market orientation are the gathering of market information, sharing the information with others in the organization and responding to market information to satisfy customers. Within the proposed internal customer orientation framework, this would entail gathering internal market information regarding internal customers needs and identifying how satisfying those needs would help the internal customers be able to better satisfy external customers needs. Sharing market information would mean, 1) sharing the internal market information across relevant internal suppliers, and 2) sharing the information across internal customers. Finally, employees must be motivated to respond to the internal market information in a way that ultimately facilitates the satisfaction of the external customer. Although this may appear to be an arduous task for employees, keep in mind that employees would only need to be gathering and responding to information relative to their own internal and external employees and not employees across the entire organization. Central to the proposed framework is the difculty in attempting a dramatic change in all aspects of employee behavior and attitudes. For an organization that is attempting to move from a low degree of market orientation to a high degree of market orientation, expecting instant change in employee behavior is unrealistic. Top management must communicate the desired market-oriented culture, and this must be done on a continuous basis (Day 1994; Jaworski and Kohli 1993). However, top down communication is not sufcient to bring about the desired change in employee behavior. A management process is required that is structured to bring about this change in employee behavior. Based in part on the work of Porras and Robertson (1990) and Garver and Cook (2001), the proposed framework suggests the changes required of employees should occur in a series of managed steps or phases that gradually shape employee behavior into the desired market oriented behavior. Changing the behavior of employees by changing the structure of the performance management system is seen as the ultimate objective of this framework. To do this, employee knowledge and attitudes regarding a market oriented culture must change, and then the organization must provide the structure to change employee behavior.
Proposed Framework For Creating a MarketOriented Culture
For a traditional organization in which the marketing function is relatively separate from other functional areas, developing a market-oriented culture can be a daunting task. To facilitate the development of a market-oriented culture, this framework breaks the development process into the two components discussed earlier (attitudes and behavior) that reect the two separate phases of the market orientation construct (Figure 2). Phase I of this framework focuses on three steps designed to change employee knowledge and attitudes to better reect a market-oriented culture. Phase II outlines three steps for changing employee behavior in a way that will create and maintain a market oriented environment. Phase I: Educating the Employees: Changing Worker Knowledge and Attitudes Employees will not change their behaviors in the work place until there has been a clear explanation of what it is management is trying to accomplish. The more support
Figure 2 Six-Step Process Model
PHASE ONE: Educating Employees: Changing Employee Knowledge and Attitudes Step One Step Two Step Three Employee education on the need for internal and external customer focus Employee education regarding the dissemination of market intelligence Communicating a system to reward responsive behavior
PHASE TWO: Implementation of a Performance Management System to Change Employee Behavior Step Four Step Five Step Six Encourage gathering of market information for every employee Encourage dissemination of market intelligence Rewarding responsive behavior
49
Martin and Martin
there is for market-oriented activities at the top of the organization, the stronger the message to the work force will be. Accomplishing Phase I is divided into the following three management steps. Step 1: Employee education regarding the need for both an internal and an external customer focus. Harris (1999) suggested that employees are potential barriers to creating and/or sustaining a market orientation for a variety of reasons. Employees tend to have a short-term perspective that runs counter to the longer-term focus within the market orientation construct. Employees tend to focus their efforts on their own productivity rather than on how their productivity benets others in their organization. Most employees overlook the fact that their efforts are connected to the organizations customers, negating the customer-oriented aspect of the market orientation construct. Most U.S. organizations have a corporate culture that emphasizes individualism and productivity. Employees understand their job to be generating as many units of work as they are able to produce. This form of behavior and its accompanying attitudes are typically reinforced through a performance management system that compensates employees for their productivity in terms of work units produced. Because a central tenet of the market orientation concept is a customer focus (Slater and Narver 1999), the key for initiating the development of a market orientation is to refocus the individualism and productivity aspects of an organizational culture toward internal and external customer satisfaction. The organization should begin with a series of department-wide seminars that clearly explain the concepts of internal and external customers and each workers role within the supplier/customer dyad. It is important to recognize that each employee would be involved with information to and from only his/her own customers and not the entire company. To facilitate this education process, the organization can map out the series of dyads that exist between employees in the organization, indicating each employees role as an internal supplier and an internal customer. We refer to this map as a dyadagram. A dyadagram is a series of mapped dyads that indicate the supplier/customer nature of the relationship between pairs of individuals in an organization. The dyadagram originated from a strength-of-ties perspective evolving out of social network theory (Uzzi 1996; Marsden and Campbell 1984; Granovetter 1973, 1982). The strengthof-ties perspective focuses on the sharing of many types of information between social actors in a social network. The dyadagram is more specic than this because it refers to an actual mapping of a series of work relationships and is restricted to internal customer-supplier relationships. Each individual in a rm has at least one dyadic relationship where s/he is either the supplier of something (for example, work product) or a customer/receiver of something. In many instances an individual worker will be involved in multiple dyadic relationships functioning as a supplier for specic individuals and a customer with others. The dyadagram focuses solely on customer/supplier relationships within 50
an organization. Applied to the development of market orientation through internal supplier/customer dyads, the dyadagram would be able to guide the employee to those supplier/customer dyads that are more or less critical to different aspects of the employees job. The dyadagram will allow employees across different departments to visualize their roles within their own close networks and their role in the organizations relationship with external customers. To strengthen the point that employees are both suppliers and customers of others in the organization, top management should maintain and communicate artifacts, or organizational stories of critical incidents that are examples of internal supplier/customer relationships that ultimately improved external customer satisfaction. Such cultural artifacts are an important means for reinforcing the marketoriented culture that top management is pursuing (Homburg and Pesser 2000).
. . . top management should maintain and communicate . . . organizational stories of critical incidents that are examples of internal supplier/customer relationships that ultimately improved external customer satisfaction.
Once employees know who their own internal customers and suppliers are, the second part of this process includes the identication of the specic internal and external market information that is required or expected to be collected and maintained by employees for each job and the sources of that information. The dyadagrams would indicate those within the organization from whom information should be collected and maintained by the employee. The market information that is required would include information about the employees internal customers and the companys external customers. For those dyadic relationships between internal suppliers and customers that are identied as stronger, the employee might be expected to collect and maintain information that is more in-depth about the internal customers needs and how those needs relate ultimately to the external customers needs. For those dyadic internal relationships that are identied as weaker, the employee might expect to collect and maintain a broader variety of information, but with less depth (e.g., Rindeisch and Moorman 2001). In addition to the need-related information for those internal customers in weaker dyads, the broader variety of information may also provide the employee with a broadened perspective on his/her job and role within the organization as well as new insights and ways to creatively satisfy internal customers so that they can satisfy external customers. Finally, for some positions within the organization for which there is little external customer contact, there may still be opportunities to gain external market information that would help satisfy internal customers who must satisfy external customers. The organization should encourage employees in these positions
Martin and Martin
to identify these opportunities and to collect the market information the opportunities may provide. Step 2: Employee education regarding the dissemination of market intelligence. That employees at varying levels of the organization will possess a clear understanding of what should happen to the market information generated in Phase I is an unrealistic assumption. Management must decide on the most efcient means for educating employees to their part in the dissemination process. While many options exist to accomplish this task, the goal is for employees to share their knowledge of internal and external markets with others in the organization. Relying on the previously described dyadagrams, focus would be placed on the most appropriate linkages for disseminating the information. Following a strength-of-ties perspective (Marsden and Campbell 1984), the dyadic linkages would help to identify with whom the information should be shared as well as what information should be shared. To facilitate this education process, whole plant or large unit seminars may work well in some instances, where small department meetings may be better for groups less familiar with the market orientation concept and the sharing of market information. In most organizations there will likely be a combination of written materials and group meetings scheduled so that all employees have a full understanding of information dissemination. Regardless of the process employed, the main objective of this step is to clarify with the organizations employees why the dissemination of market intelligence is so important. The employees need to recognize that gathering information is important, but that the information needs to be shared so that it can be acted upon by various individuals or departments. Step 3: Communicating the reward system to encourage responsiveness to market information. The nal part of the rst phase is to close the loop in the employees understanding of the new system. Thus far the workers have received an explanation of the need for an internal and external customer orientation and their role in generating and disseminating both internal and external market information. At this point employees need to understand what the company would like for them to do with the market information they receive. Some companies will encourage workers to act independently and make changes themselves that will positively affect the consumer. Other companies prefer ideas be discussed between the workers and their supervisor within a workers own department or manufacturing cell, and then acted upon, while still other companies prefer all ideas be submitted for approval to higher levels of management before being acted upon. Critical to this process is that whatever system the company chooses, it must be clearly communicated to employees. In addition, it is crucial for top management to determine how the processes of intelligence generation, dissemination and responsiveness will be tied into the performance evaluation of the work force. Whether they use the balanced scorecard approach of Kaplan and Norton (1992) or a more traditional performance management system, it is important
that the manner in which market-oriented behavior will be tied to rewards is explained to everyone in the organization. The reward system is a key element of Phase II and will be discussed in more detail subsequently. In summary, Phase I of the system being proposed is focused on education of the workforce. From the top to the bottom of the organization the employees need to know that management is committed to gathering, disseminating and responding to internal and external market information and that employees are expected to engage in these behaviors to a certain extent.
...it is crucial for top management to determine how the processes of intelligence generation, dissemination and responsiveness will be tied into the performance evaluation of the work force.
Phase II: Implementation of a Performance Management System to Change Employee Behavior Step 4: Encourage the gathering of market information as part of every employees job. The fourth step in the process requires employees to go from the acknowledgement that a customer focus is important, to changing their behavior to support a customer focus and the generation of market information. Improving knowledge and attitudes of employees is insufcient to change their performance. An additional, necessary means for encouraging the generation of market intelligence is through a performance management system. A performance management system can play a pivotal role in inuencing organizational behavior because a performance management system is a network of related components whose ultimate purpose is to improve organizational effectiveness (Beer, Ruh, Dawson, McCaa and Kavanagh 1978). A performance management system is an initiative proposed by top management that sets up a process for on-going evaluation of worker productivity along with continuous feedback, with the ultimate goal being continual development of employee skills and activities that enhance the effectiveness of the rm. A performance management system is designed to improve an organizations effectiveness and efciency through changing the behavior of its employees. It requires cooperation from all levels within the organization. The strategic and operational goals of the rm should be developed with a market-oriented culture being a key component. In most organizations employees have generally been rewarded for quantity and quality of production. Their job descriptions are clear about detailing the specic activities of their jobs. The fourth step in developing a market orientation involves re-writing the job descriptions and performance standards to include the actual activities that are necessary for generating market intelligence. The behaviors necessary to generate the market information would be included in the job description and yearly objectives for each position. 51
Martin and Martin
The performance management system would include rewarding behaviors oriented toward generating market intelligence. This would involve detailing a reward program that will entice employees to allocate a certain percentage of their time and effort to generate market intelligence. All workers have a limited amount of time and energy that must be divided between the various aspects of their jobs. Motivating employees to allocate a portion of their resources to generating market intelligence through the reward system is a necessary component in the performance management system. The organization must not only clarify the reward system, but must also identify the nature of internal customer information the employee is expected to collect. This information would focus on the specic needs that the internal customer has in order to successfully perform his/her job. These needs will vary depending on the situation and the nature of the relationship between supplier and customer. Given the plethora of possibilities regarding the type of internal customer information the employee might collect, it is beyond the scope of this paper to provide any sort of detailed specication. However, we believe a fruitful area of research might pursue the development of a taxonomy of internal customer information. As an example of how the rst four steps would t together, consider the job of a product design engineer. Prior to initiating a market orientation in a company, the design engineers focus would probably be the design of a product to meet product specications. The design engineer would not necessarily be concerned with satisfying the needs of operations people or of the product or sales staff for that product and, therefore, would design the product from an efciency and engineering perspective. Once top management becomes committed to changing the organizations product design culture to a market-oriented culture, the rst initiative is to develop a dyadagram for this position. To do this one must identify the internal supplier/customer dyads that include the design engineer. For example, the engineer is, at
a minimum, a supplier to production-operations people and to product managers or sales managers for different products being designed. The engineer is also, at a minimum, a customer of the concept development team and the R & D scientist/engineers who are supplying new product concepts and ideas. The dyads could be diagrammed as in Figure 3. Also in this rst step would be the identication of the sources of information and the types of information the design engineer should collect and maintain. For example, based on the dyadagram in Figure 3, in addition to the basic product specications, the design engineer might nd that the internal operations customer may want certain aspects of the product to be designed a certain way. At the same time, the product or sales internal customer for that product may also want specic aspects of the product to be designed a certain way. Both of these internal customers may be basing their design needs on what will help them achieve better performance. Both of these internal customers may also have design needs that conict with each other. In addition, detailed market information from the concept development team or the R & D scientist may suggest alternative design possibilities not indicated in the original specications. The dyad relationships in Figure 3 suggest the design engineer must come to understand that he/she should share information from each of his/her internal customers and suppliers with his/her other internal customers and suppliers. So, for example, knowing that operations may want X and sales may want Y could be shared with R & D, which might affect subsequent versions of the product. To motivate the hypothetical design engineer in Figure 3 to collect the relevant market information, the evaluation system and the reward system would need to be structured so that the engineer is rewarded for collecting market information from all relevant sources. In the event of potentially conicting needs of different internal customers, the engineer must also be rewarded for collecting information about the external customers needs for this product from sources
Figure 3 Hypothetical Dyadagram for a Product Design Engineer

Product Design Engineer
Concept Development Team Internal Suppliers R & D Scientist
Production Operations
Other Internal and External Customers
Product/Sales Manager for Product A
Product/Sales Manager for Product B
Other Internal Customers
Other Internal and External Customers
52
Martin and Martin
identied in the dyadagram as well as sources external to the company. The precise nature of the reward system must be clearly communicated to the design engineer. For example, if the engineer attends a professional conference, there is a potential source available for generating additional market information. However, the engineer is also involved in attending formal presentations, meeting with colleagues about advances in the eld and nding time for rest and relaxation. Each of these activities has obvious rewards for the engineer. If a company wants the engineer to further sub-divide his/her time to talk with competitors or customers about product development, the company must reward such behavior. This could be done through a variety of mechanisms. For example, the conference stay could be extended one day in order to give the engineer more time for generating market information or there might be individual departmental incentive programs. Regardless of the specic reward mechanism used, the criteria for performance (in this case the collection of relevant information) must be clear and the reward must be of value to the engineer. The process of gathering and sharing internal customer information and external customer information would be streamlined by the use of a computer software system that would store and disseminate market intelligence based on a dyadagrammatic design of the rm. Then as individuals collected market information they could store and disseminate it through a central location within the companys internal computer network. As discussed by Porras and Robertson, 1990; (see Figure 1), this would be an important change in technology that would result in individual behavior change. The last two processes describe the steps necessary for employees to share internal and external market information and to respond to that market information. Both steps may occur at the same time, within a short period of time, or each step may require lengthier periods of time for implementation, depending on the organizations current culture. Step 5: Encouraging dissemination of market intelligence. Once market intelligence is generated there is the need to disseminate the information to relevant parties within the organization. There can be no response to the information without it rst being communicated throughout the organization. The main focus in this step is to motivate those individuals who have accumulated market information to take the time and effort to disseminate their information across the organization. Because people tend to perform those activities for which they are rewarded, the company must set up a system for rewarding the dissemination of market intelligence. This step raises several issues about dissemination of information. For example, should the dissemination be formal or informal, how much information should each employee receive, should employees receive sensitive customer or nancial information, how should a reward system be structured for sharing information, how often should employees be expected to disseminate information and how often dissemination should be measured are just a few of the
issues managers must grapple with in their attempt to implement a market orientation. Because of the wide variation within organizational and market environments, no rules are currently available to give easy answers to these questions. Generally, wider dissemination of information is considered better for an organizations ability to respond to that information. There will, however, be an array of exceptions to this. Customer and nancial information is often shared with employees in highly market-oriented rms, but we are sure there are situations where this is not the case. Reward systems are structured very differently across rms.
Whether dissemination should be formal or informal is also idiosyncratic to the organizations culture.
Whether dissemination should be formal or informal is also idiosyncratic to the organizations culture. For example, in Figure 3, if the manager for product A has received external market information that the customer is dissatised with the placement of an on/off switch on a particular product, then the product/sales manager needs to communicate that information to the product design engineer. The rm should develop a reward structure that will reinforce the product manager for taking his/her time to pass the market information along to the product engineer. Whether this is done in a formal report or during an informal conversation would depend on the organizations culture. As another example of a dissemination method, one consumer products company developed and circulated a newsletter to disseminate market intelligence. Contributions to the newsletter were an active part of the job description. The company set up an incentive system for contributions to the newsletter to encourage employees across the organization to contribute. For another company, a more efcient possibility might be an internal web site dedicated strictly for market intelligence that would allow for a full dissemination of information across all levels of the organization. Employees job descriptions can include expectations for the frequency of visiting the site and incentives can be offered for contributions to the site. Regardless of the specic mechanism used, as employees begin more and more horizontal and vertical dissemination of market intelligence, the company should expect a shift in employee attitudes to reect a market driven culture at all levels of the organization. Step 6: Responsiveness to market intelligence: Rewarding responsive behaviors. Simply having market information available within the organization accomplishes little. The organization must respond to the information in a way that provides a competitively superior value to customers in a timely manner. A company can be responsive in several ways from re-designing products to offering new products, changing the distribution and promotion of products to 53
Martin and Martin
changing servicing of products or of customers. By developing a performance management system that reinforces these behaviors the company will establish commitment at all levels of the organization.
The organization must respond to the information in a way that provides a competitively superior value to customers in a timely manner.
By phase II all workers should be educated to understand the customer orientation of the organization. They should have available market intelligence so that now it becomes imperative for a company to establish a system that encourages the employee to respond to the market information. For example, in a retail store if information generated from customer comments indicated that external customers found it distressing when clothes were disorganized on the shelves, then the company might want its sales clerks to spend more time straightening the merchandise on the racks and shelves. However, if sales clerks are paid entirely on commission there is little motivation to spend time straightening up the clothes. Therefore, evaluating and compensating this employee based on activities that are responsive to the market information they have received is critical. Consider the product design engineer in Figure 3. The engineer has now collected information from the product manager suggesting that external customers dont like the placement of an on/off switch. The engineer shares that information with production/operations people who like the switch where it is because it is less costly to attach in that spot. The organizations reward system must be set up such that the engineer receives a greater reward if a solution is found that uses both pieces of information thereby satisfying both sets of internal customers and ultimately the external customer. While the response to market information is a necessary step, that response should be measured for its appropriateness. If the product design engineer changes the position of a switch in response to intelligence received, the appropriateness of the change should be determined before the reward for responsiveness is offered (Hauser, Simester and Wernerfelt 1996). Efforts by Garver (e.g. Garver and Cook 2001; Garver and Gagnon 2002) suggest that customer satisfaction (for both internal and external customers) is the appropriate metric to use to evaluate responsiveness. As those authors point out however, customer satisfaction must be balanced against the protability of achieving that satisfaction and this must also be taken into account. Several implementation issues emerge as the manager works through this nal step. In addition to the managers development of training modules for educating employees about the importance of market-oriented behaviors (Phase I), implementation of the six steps requires training to develop the skills necessary to engage in market oriented behaviors (Phase II). Employees recognizing the impor54
tance of market oriented behaviors will accomplish little if employees are not trained to perform the behaviors. Every company and every type of position will require a unique variation of training, and although it is beyond the scope of this paper to develop individual training modules for each employee position, there are several tools in the training literature and the TQM literature that could be helpful to a manager in the implementation process. For example, based on the TQM literature, the rst and perhaps most important aspect of implementation would be for top management to be directly involved in some way with the implementation process. Demings management-by-walking-around approach is an excellent method that involves top management and will also communicate top managements vision and commitment to a market oriented culture. Cause-and-effect diagrams and root cause analysis, check sheets and data collection sheets, graphic displays of data and processes, and ow charts for process mapping are excellent methods for helping an employee learn how to collect information pertaining to the links between his/her actions and his/her internal customers needs. Other tools to assist the employee with the acquisition of information could include Taguchis loss function approach which would focus the employees attention on the variation of his/her behavior around a targeted goal for his/her internal customers, self audits, who-what-where-when-why-how analysis, and evolutionary operations analysis. Tools to help employees learn ways to be responsive could include a plan-do-check-act cycle, brainstorming, a root cause analysis, and goal setting. QFD matrices and force eld analysis are tools that can help the employee translate his/her internal customers needs into effective responsiveness. Most books on TQM describe these tools in detail (e.g., Brocka and Brocka 1992; Crosby 1979; Deming 1982; Hodgetts 1996; Juran 1988; Pegels 1995). Other implementation issues the company must also consider include how responsiveness will be measured and whose task it will be to track this information. Finally, in addition to collecting and tracking the performance information, decisions must be made regarding the reward structure. Will rewards be provided on a weekly, quarterly or yearly basis? When responses are still uncertain and being learned it is best to reward desired behaviors as frequently as possible. Therefore, for the rst year of this plan it would be useful to plan feedback and rewards on a monthly basis, perhaps cutting back to quarterly over time. For on-going behavior, a yearly review is likely to sustain the desired level of responsiveness. However, the timing of rewards is also contingent on the frequency of opportunities for responding to market information. In addition, the organization must provide detailed training that will show employees how to use internal and external customer information. As rms increasingly recognize the strategic importance of becoming more market oriented, organizations will necessarily focus on the problem of driving a market orientation deep into the organizations culture. This focus brings the organization face-to-face with the barriers and difculties
Martin and Martin
of implementing a market orientation. Central to those difculties is the necessity of changing employee behavior to reect the different orientation of the organization.
Company Example
The following is an example of how one company used the above framework to implement an organization-wide market orientation. Cardinal Fastener (40 employees; less than $20 M in revenue) is a small manufacturing company that produces fasteners used in the construction of OEM equipment such as domes, stadiums, oil rigs, and drilling cranes. When Cardinal was purchased several years ago one of the main goals of the new owner/president was to grow the company and to establish a dominant position in a national market. The company already had an established name for quality and turned its attention to focusing on the customer. The rst phase of implementing a market orientation is educating the work force. The new president began by explaining to his work force the concept of internal and external customers. Next he physically lined up the employees across the plant to show them their own internal suppliers and customers giving employees a very personal sense of the dyadic relationships in which they were involved. Each worker stood arm-to-arm with his/her own customers and suppliers as the president explained to the workers why knowing their internal customers and suppliers is important information. He had several employees stand up and talk about their internal relationships so that others would begin to understand, thus establishing cultural artifacts (Homberg and Pesser 2000) that would continue to reinforce the internal customer-internal supplier concept. Once the workers knew who their own internal customers/suppliers were, the second and third processes in Phase I took place; the nature and type of market information they were to gather was explained, as was the reward structure. The reward would be simple and straightforward; they were to be given weekly cash awards for solving problems and having error-free ontime delivery. These goals were linked to the specic performance of each job through the internal customer internal supplier network throughout the organization. In this way each employee knew exactly how his/her behavior affected his/her internal customers and how this led to satisfaction of external customers. The second phase of the process for Cardinal was changing employee behavior. While the initial organization-wide meetings were the most difcult, once the employees began to understand the terminology and how the system worked, the president felt that the change began to generate its own momentum. To this day there continue to be monthly company-wide meetings where attendance is required. The plant shuts down for about forty-ve minutes while the past months performance is evaluated and plans for the following month are discussed. This process accomplishes two elements discussed previously. It provides the opportunity to discuss market information and it also communicates
the importance of maintaining a market oriented focus throughout the company. Each meeting focuses primarily on a single issue such as delivery, quality or competitors. The president tries to encourage every employee to share information with each other. Every individual in the organization is encouraged to share in these discussions. One way this is done is to use silver dollars. When an individual shares market information with the rest of the workers he tosses them a silver dollar. The silver dollar is obviously a minimal monetary incentive, but the president of Cardinal has found that it acts as a very strong social reward. Although a small token, he has found this method to be very motivating in encouraging the workers to share market information. Beyond the additional monthly monetary incentives, workers are praised and supported for gathering and sharing market information with one another.
Beyond the additional monthly monetary incentives, workers are praised and supported for gathering and sharing market information with one another.
Top management takes an active role in sustaining the market orientation within Cardinal Fastener. Market information continues to be shared on an on-going basis. Top management practices dissemination of market information by sharing both competitor and customer information company wide. The top customers and target customers are posted on bulletin boards around the plant. Delivery schedules are posted daily. An example of responsiveness to market information is their program called Golden Nuggets. Each problem or new idea is called a Golden Nugget. This emphasizes the point that problem identication is good for the company, because only then can the problem be solved. Each problem is turned into an opportunity. Emergence of problems is considered to be the result of one of two reasons: either a procedure was not followed, or there was no procedure to follow. When a problem is identied all of those individuals responsible meet with all of the internal customers and suppliers in their chain to analyze the problem and to generate a solution. As an example, there was an order that was improperly lled. The problem was identied as an error in transcribing what the customer requested onto the job order form. The standard procedure had been to proof read each order carefully, but this had failed to produce an error-free result. The solution was that the individuals taking the orders would now use multi-colored highlighters in order to cross check line by line that each order was being transcribed correctly. Another interesting example focused on the shipping department. An external customer complained that one of their orders had the wrong number of pieces in the carton. The carton was to hold twenty-ve pieces but had arrived with only twenty-four pieces. Previously, the shipping department packaged the items into open boxes for shipping. To 55
Martin and Martin
solve the problem the sales and shipping employees got together along with the team leader in the manufacturing cell (their internal supplier). The solution was to have the pieces packed in the manufacturing cells using a subdivided tray with twenty-ve clearly visible compartments. The result was that the external customer was happy and a step was cut in the manufacturing process. Thus, the Golden Nugget generated information to be shared among internal customers and suppliers and resulted in responsiveness to the problem and to the external customer. In both of the above examples employees are treating each other as internal customers and suppliers, but always with the ultimate goal of serving the external customer. Cardinals president feels that employees need to be given every opportunity to excel. Trying to change a corporate culture is difcult. Once in place, sustaining it is less taxing, but it requires a continuous focus on the issues of internal and external customers. The workers see each other as customers and treat each other with respect, ultimately increasing both speed and efciency of production for the external customer. The result of driving the market orientation deep into the organization has been a tremendous growth in the company (approximately 20 percent annual growth for the past ve years), high quality products, competitively superior delivery time, extremely high worker satisfaction, lower costs and higher prots.
What makes this process ideal for small organizations is the ability of the company president to become personally involved in the process. The president can directly communicate with employees to emphasize the connection between specic work behaviors and specic organizational rewards. In addition, dyadagrams are easily developed and understood in a small organization. Workers can readily grasp the relationship between their work and that of their internal customers and suppliers and how this affects the external customer. Future research might be focused in three separate areas. One potential area would be developing a theoretical model around the concept of the dyadagram. The intent would be to lead to organizational designs of the internal network and technical systems for uid communication within the network. A second area for future research is to identify which of the proposed six processes is currently being used by organizations and how might those steps be improved upon. Investigating which management techniques are most inuential to the process and which metrics are most benecial for measuring the effectiveness of the change would be additional important areas for future research. A nal area for research is to study the effects of the reward structure to determine how best to tie rewards to increased intelligence generation, dissemination and responsiveness.
References Conclusions and Future Research

Despite the general acceptance of the importance of a marketing orientation to an organizations protability, to date little has been offered that aids businesses as they undertake such an endeavor. The main objective of this paper was to offer six management steps within two separate phases that would provide a framework for initiating a market orientation. The six steps rely on an internal customer internal supplier orientation as the link between the organizations employees and its external market. Following the suggestion of Hauser, Simester and Wernerfelt (1996), the proposed implementation structure links satisfaction of internal customers to the ability of the organization to satisfy external customers. The dyadic approach establishing stronger and weaker ties between internal customers and suppliers provides the foundation for changing the organizations performance management system in a way that shapes employee behavior toward the desired market oriented culture. As is clear from both the framework and the example with Cardinal, managers can incorporate techniques with which they are familiar into the process. Such techniques as root cause analysis, QFD, TQM, and process mapping are a few that could be investigated for their roles in enhancing the implementation of this process. The nature of how tools such as these would t into the framework should be explored. 56
Achrol, R., and P. Kotler. 1999. Marketing in the network economy. Journal of Marketing. 63 (Special Issue): 146-163. Anderson, E. and R. L. Oliver. 1987. Perspectives on behavior-based versus outcome-based salesforce control systems. Journal of Marketing. :76-88. Appiah-Adu, K. 1997. Market orientation and performance: Do the ndings established in large rms hold in the small business sector? Journal of Euro-Marketing. 6:1-26. Appiah-Adu, K., and A. Ranchhod. 1998. Market orientation and performance in the biotechnology industry: An exploratory empirical analysis. Technical Analysis and Strategic Management. 10:197-210. Avlonitis, G., and S. Gounaris. 1997. Marketing orientation an company performance. Industrial Marketing Management. 26:385-402. Baker, T. L., P. M. Simpson, and J. A. Siguaw. 1999. The impact of suppliers perceptions of reseller market orientation on key relationship constructs. Journal of the Academy of Marketing Science. 27:50-57. Beer, M., R. Ruh, J.A. Dawson, B. B. McCaa, and M. J. Kavanagh. 1978. A performance management system: Research, design, introduction and evaluation. Personnel Psychology. 31:505-535. Brocka, B. and M.S. Brocka. 1992. Quality management: Implementing the best ideas of the masters. New York: Irwin. Cannon, Joseph P. and Christian Homburg. 2001. Buyer supplier relationships and customer rm costs. Journal of Marketing. 65(1):29-43. Cardy, R. L. and T. T. Selvarajan. 2001. Management Interventions. In Handbook of Industrial, Work and Organizational Psychology, Ed. N. Anderson, D. Ones, H.K. Sinangil, and C. Viswesvaran. Thousand Oaks, CA: Sage Publications.
Martin and Martin
Chatterjee, S. and B. Wernerfelt. 1991. The link between resources and type of diversication: Theory and Evidence. Strategic Management Journal. 12:33-48. Conduit, J., and F. T. Mavondo. 2000. How critical is internal customer orientation to market orientation? Journal of Business Research. 51:11-24. Crosby, P. B. 1979. Quality is free. New York: McGraw-Hill. Day, G. S. 1994. The capabilities of market-driven organizations. Journal of Marketing. 5:37-52. Deming, W. E. 1982. Quality, productivity and competitive position. Cambridge: MIT Press. Deshpande, R. (Ed.). 1999. Developing a market orientation. Thousand Oaks: Sage Publishing. Deshpande, R., and J. U. Farley. 1999. Executive insights: Corporate culture and market orientation: Comparing Indian and Japanese rms. Journal of International Marketing. 7:111-127. Deshpande R., J. U. Farley and F. Webster, Jr. 1993. Corporate culture, customer orientation and innovativeness in Japanese rms: A quadrad analysis. Journal of Marketing. 57:23-37. Garver, M. S., and R. Cook. 2001. Best practice customer value and satisfaction cultures. Mid-American Journal of Business. 16(1). Garver, M. S., and G. Gagnon. 2002. Seven keys to improving customer satisfaction programs. Business Horizons. 16(2):35-42. George, W. R. 1990. Internal marketing and organizational behavior: A partnership in developing customer-conscious employees at every level. Journal of Business Research. 63-70. Goetsch, D. L. and S. B. Davis. 1997. Introduction to total quality. Upper Saddle River: Prentice Hall Granovetter, M. 1973. The strength of weak ties. American Journal of Sociology. 78:1360-1380. Granovetter, M. 1982. The strength of weak ties: A network theory revisited. in Social Structures and Network Analysis. Ed. P.V. Marsden and N. Lin. 201-233. Beverly Hills:Sage Publications Grewal, R., and P. Tansuhaj. 2001. Building organizational capabilities for managing economic crisis: The role of market orientation and strategic exibility. Journal of Marketing. 65(2):67-80. Gronroos, G. 1990. Service management and marketing: Managing the moments of truth in services Competition. Lexington, Mass: Lexington Books. Hackman, J. R. 1981. Sociotechnical systems theroy: A commentary. In Perspectives in Organization Design and Behavior. Ed. A. H. van de Ven and W. F. Joyce. New York: Wiley. Hallowell, R. L., A. Schlesinger and J. Zornitsky. 1996. Internal service quality, customer and job satisfaction: Linkages and implications for management. Human Resource Planning. 19(2):20-31. Han, J. K., K. Namwoon, and R. K. Srivastava. 1998. Market orientation and organizational performance: Is innovation a missing link? Journal of Marketing. 62(4):30-45. Harris, L. C. 1999. Barriers to developing market orientation Journal of Applied Management Studies. 8:85-91. Hauser, J. R., D. I. Simester, and B. Wernerfelt. 1996. Internal customers and internal suppliers. Journal of Marketing Research. 33:268-280. Heilmann, R.L. 1994. The alignment matrix. Quality Progress. 27:75-78. Hodgetts, R. M. 1996. Implementing TQM in small and medium sized organizations: A step-by-step guide. New York: Amacom Publishing. Homburg, C., and C. Pesser. 2000. A multiple-layer model of market-oriented organizational culture: Measurement issues
and performance outcomes. Journal of Marketing Research. 37:449-462. Hooley, G. J. L., J. Lynch., and J. Shepherd. 1990. The marketing concept: Putting the theory into practice. European Journal of Marketing. 24:7-23. Hunt, S. D. and R. M. Morgan. 1995. The comparative advantage theory of competition. Journal of Marketing. 59:1-15. Jaworski, B. J., and A. K. Kohli. 1993. Market orientation: Antecedents and consequences. Journal of Marketing. 57:53-70. Juran, J. M. 1988. Quality control handbook. N.Y.: McGraw-Hill. Kaplan, R. S. and D. P. Norton. 1992. The balanced scorecardmeasures that drive performance. Harvard Business Review. (January/February):71-79. Kohli, A. K., and B. J. Jaworski. 1990a. Market orientation: The construct, research propositions, and managerial implications. Journal of Marketing. 54:1-18. Marsden, P. V., and K. E. Campbell. 1984. Measuring Tie Strength. Social Forces. 63:482-501. Matsuno, K. and J.T. Mentzer. 2000. The effects of strategy type on the market orientation performance relationship. Journal of Marketing. 64(4):1-16. Mohr-Jackson, I. 1991. Broadening the market orientation: An added focus on internal customers. Human Resource Management. 30:455-467. Narver, J. C., and S. F. Slater. 1990. The effect of market orientation on business protability. Journal of Marketing. 54:20-35. Pegels, C.C. 1995. Total quality management: A survey of its important aspects. New York: Boyd & Fraser Publishing. Pelham, A. M. 2000. Marketing orientation and other potential inuences on performances in small and medium-sized rms. Journal of Small Business Management. 45-67. Pelham, A. M. and D. T. Wilson. 1996. A longitudinal study of the impact of market structure, rm structure, strategy and market orientation culture on dimensions of small-rm performance. Journal of the Academy of Marketing Science. 24(1):27-43. Porras, J. I., and R. C. Silvers. 1991. Organizational development and transformation. Annual Review of Psychology. 42:51-78. Porras, J. I. and P. J. Robertson. 1990. Organizational development: Theory, practice and research. In Handbook of Industrial and Organizational Psychology. Ed. M. Dunnette, and L. Hough. Palo Alto: Consulting Psychologists Press. Porter, L. W., and E. E. Lawler. 1968. Managerial attitudes and performance. Homewood: Irwin. Rindeisch, A. and C. Moorman. 2001. The acquisition and utilization of information in new product alliances: A strength-of-ties perspective. Journal of Marketing. 65(2):1-18. Siguaw, Judy A., P. M.Simpson., and T. L. Baker. 1998. Effects of supplier market orientation on distributor market orientation and the channel relationship: The distributor perspective. Journal of Marketing. 62:99-111. Slater, S. F., and J. C. Narver. 2000. The positive effect of a market orientation on business protability: A balanced replication. Journal of Business Research. 48:69-73. Slater, S. F., and J. C. Narver. 1999. Market-oriented is more than being customer led. Strategic Management Journal. 11651168. Slater, S. F., and J. C. Narver. 1994. Does competitive environment moderate the market orientation-performance relationship? Journal of Marketing. 58:46-55. Steinman, C., R. Deshpande, and J. Farley. 2000. Beyond market orientation: When customers and suppliers disagree. Journal of the Academy of Marketing Science. 28:109-119.
57
Martin and Martin
Uzzi, B. 1996. The sources and consequences of embededness for the economic performance of organizations: The network effect. American Sociological Review. 61:674-698. Wernerfelt, B. 1989. From critical resources to corporate strategy. Journal of General Management. 14:4-12. Wrenn, B. 1997. The market orientation construct: Measurement and scaling issues. Journal of Marketing Theory and Practice. 5(3):31-54.
In Kalamazoo and Grand Rapids, MI About the Authors

Beth Ann Martin is Professor of Industrial/Organizational Psychology at John Carroll University. Her research interests focus on the development of linkages between performance management systems and market orientation in small and mid-sized organizations in addition to the use of personality assessments in employment selection decisions. martin@jcu.edu James H. Martin is Professor of Marketing in the Boler School of Business at John Carroll University. His research interests focus on strategies that enhance the development and implementation of a market orientation, especially comparing rms in transition or developing economies to rms in Western developed economies. jhmartin@jcu.edu
The convenience you need. The expertise you expect...

Our EVENING MBA PROGRAM is designed around your needs and expectations. The MBA curriculum, faculty and students will support your career objetives, while the administration and course scheduling will support your busy lifestyle. Our cross-functional approach to your MBA will help you solve real business problems and you will view your business from new perspectives. You will develop critical analysis and other business skills to help you advance your career while establishing a network of friends and colleagues with whom youll exchange ideas for years to come.
Invest in your future...
come to Western Michigan
then go far.
Call: 269-387-5075 Email: bus-adv-ofce@wmich.
58
Implementing Planned Change:
An Empirical Comparison of Theoretical Perspectives

Matthew W. Ford, Northern Kentucky University Bertie M. Greer, Northern Kentucky University
Abstract
Planned organizational change has been viewed from a variety of conceptual perspectives, and a plethora of variables that impact the change process have been proposed. However, few empirical studies have investigated the relationships thought to exist among change process variables. Drawing from questionnaire-based data obtained from managers involved in the implementation of change, we evaluate three plausible change model congurations using multivariate methods. Findings from the study support a dynamic change process conguration over a direct effects model. Results, discussion, implications and direction for further research are offered.
...few models of planned change have been studied using empirical research designs.
Introduction
Understanding the process of planned change is imperative for managers who are charged with implementing strategic initiatives that drive the success of the organization. Planned change refers to a premeditated, agent-facilitated intervention intended to modify organizational functioning for a more favorable outcome (Lippit, Watson, and Westley 1958). This perspective largely reects the teleological category of change theory advanced by Van de Ven and Poole (1995) in which organizational change is achieved primarily through the adaptive behavior of individuals in light of internally set goals. While other perspectives stress the role of external, Darwinian-like forces in organizational change (e.g., Alchian 1950; Hannan and Freeman 1977), a substantial volume of literature favors the teleological premise of premeditated actions to effect change (Huy 2001). Further, the notion of crafting and deploying large-scale change initiatives has been widely diffused among managers as the basis for strategic management (e.g., Andrews 1971; Thompson and Strickland 1998). Although planned change has been viewed from a variety of conceptual perspectives (e.g., Gioia and Chittipeddi 1991; Huy 2001; Levy 1986), few models of planned change have been studied using empirical research designs. Inquiry using empirical methods could illuminate a number of issues about which we remain largely uninformed such as the
relative importance of various change process factors in successful change implementation. For example, is incentive system alignment more important to change achievement than, say, skill development and delivery? Empirical studies could also assist in more accurately specifying change process models. Although many change process congurations have been proposed, few have been tested to determine how factors should be organized to best express the process of change (Pettigrew, Woodman, and Cameron 2001). Some resolution could be obtained by testing competing model congurations with empirical data, and linking the resulting measurement properties to the models appropriateness (Venkatraman 1990). In this study, we seek to contribute to both theory and practice by investigating the process of planned change in an empirical context. First, we extract common factors from several prominent conceptualizations of planned change process. We then use these factors to congure three alternative models of change process implied by the literature. Using data gathered from over one hundred managers involved in the implementation of planned change, we employ multivariate methods such as factor analysis and structural equations modeling to evaluate and compare the models. By evaluating the construct and predictive validity of the models, we draw conclusions about the appropriateness of the three plausible congurations, and about the relative importance of various change process factors in achieving implementation success.We conclude by discussing the practical implications of the studys ndings and by offering further direction for change process research.
Theoretical Background and Hypothesis Development
Change Variables Planned change is often conceptualized as a process because of the sequence of actions or events that unfold to 59
Ford and Greer
move the organization from one state to another (Garvin 1998). Models of change process tend to share three basic stages (Kanter, Stein, and Jick 1992). The rst stage involves questioning the organizations current state and dislodging accepted patterns of behavior. The second stage is a state of ux, where new approaches are developed to replace suspended old activities. The nal period consists of institutionalizing the new behaviors and attitudes. These three stages are clearly visible in many classic conceptualizations of change such as Lewins (1951) unfreezing-movement-refreezing framework. The same stages can be also be used to categorize the numerous variables proposed as contributors to the change process. For example, each of Kotters (1996) eight steps for managing change are readily categorized into the various stages.
...the inuence of particular change process factors on implementation success is not equally distributed, and . . . some factors might matter more than others.
Romanelli and Tushman 1985) or continuous (e.g., Weick and Quinn 1999) in nature. The extent to which change process factors function in parallel or in sequence to produce successful change is a central issue in the literature that remains resolved. Variable Selection One way to assess the aforementioned concerns of the change process is to test competing model congurations empirically and link the resulting measurement properties to the models appropriateness (Venkatraman 1990). Since this study contributes to a relatively nascent stream of empirical change process research, we decided to limit the factors in our model to a few core building blocksfactors widely accepted as contributing to the process of planned organizational change. To obtain these factors, we studied conceptualizations of change proposed by Nadler and Tushman (1980), Tichy (1983), Burke and Litwin (1992), and Kotter (1995, 1996). These models were chosen for a few reasons. First, each of these models display some character of the teleological change theory category proposed by Van de Ven and Poole (1995). Second, these models have been widely cited in the literature; many have been featured in formal reviews of organizational change theory (e.g., Burke 1995; Werr 1995). Finally, these models represent prominent contemporary frameworks that have established a presence in the empirical world. Our comparison found ve factors common to all of these models. One factor related to activities aimed at planning or determining the organizational actions necessary to operationalize the change. A second factor reected developing and delivering new behavior to replace old patterns of action. A third factor involved aligning incentive and reward systems to encourage behavior necessary to realize successful change. A fourth factor involved monitoring of the implementation progress and taking corrective action when necessary. Finally, there was a factor that reected the change outcomes themselves, or the extent to which implementation was successful. The rst four factors became independent process variables for our investigation while the fth factor represented the dependent results variable. This small variable set allowed us to operationalize our research questions using a research design that was manageable in the present but scalable (i.e., open to the addition of more variables) for follow-up investigations. In the paragraphs below, we provide further evidence of the content validity of these ve factors. Variable Justication Action planning. Scholars have historically proposed the disaggregation of high-level goals into more concrete plans of action. Barnard (1938) argued that an organizations purpose and objectives should be broken into fragments ordered in time and assignment for cooperation. Simon (1947) portrayed an organization as a hierarchy of decisions with action at lower levels. Ansoff (1965) suggested that
Despite general acceptance of the process notion, there has been little agreement on the organizational factors or activities that comprise the process. Theorists have proposed a plethora of factors as contributing to the process of planned change. Tichy (1983) for example, proposed nine factors or levers that could be adjusted to facilitate organizational change; each factor required evaluation in the technical, cultural, and the political context of the organization. In most theoretical models, little guidance is offered about a change process factors relative importance to successful implementation. Some scholars weigh their change process factors equally and caution against ignoring any of them in the pursuit of successful change (e.g., Kotter 1995). However, there is reason to believe that the inuence of particular change process factors on implementation success is not equally distributed, and that some factors might matter more than othersat least in particular contexts. For example, it has been argued that elaborate up-front planning may hinder change achievement, particularly when the planned change is large in scale (Mintzberg and Waters 1985). The literature is also unclear on how factors reecting the process of change are best organized (Pettigrew, Woodman, and Cameron 2001). Some scholars have suggested that the process of change is sequential to some degree, and that, when implementing change, it is more important to alter some elements of the organization before others (e.g., Hinings and Greenwood 1988). Others have noted the iterative nature of planned change and its implementation (e.g., Lindblom, 1959; Quinn 1980), which challenges the notion of planned change as orderly proceeding from one phase to the next. Moreover, contextual factors may play a role in change process sequencing. The order might depend, for instance, on whether the planned change is episodic (e.g., 60
Ford and Greer
strategic objectives were best implemented through a series of cascading goals down through the organization. Plans of specic action served as linking pins between organizational levels on the way to goal achievement (Likert 1961). Action planning processes can be highly structured, particularly in the context of planning large-scale change (e.g., Hofer and Schendel 1978; Thompson and Strickland 1998). Although many changes are incremental in their development (Quinn 1980), action planning is often viewed as an early element in temporal processes of change. Skill development and delivery. Organizational change is realized largely through changes in individual behavior (Goodman and Dean 1982; Robertson, Roberts, and Porras 1993; Tannenbaum 1971), since the nature of individual behavior signicantly inuences organizational performance (Porras and Hoffer 1986). Many models of planned change emphasize the task or work related aspects of behavior change (e.g., Nadler and Tushman 1980; Weisbord 1976). Organizational change requires the development and delivery of skills in a way that will permit successful change implementation. Evidence supports the relationship between practices to acquire and develop skills and the achievement of organizational goals (e.g., Kerr and Jackofsky 1989; Terpstra and Rozell 1993). The timing of skill development and delivery must permit workers to assimilate and practice skills prior to their regular use, particularly for groups that must coordinate new skills as a work unit (Cottrill 1997). However, skills delivered too far in advance are undesirable if workers forget how to turn their knowledge into practice (Adams 1967), or if workers fail to see the connection between practicing these skills and the organizational change imperative (Baldwin and Magjuka 1997). Incentives. Incentives induce action and motivate effort (Cummings and Schwab 1973). In addition, incentive and reward systems constitute a primary governance mechanism for the organization (Jensen and Meckling 1976). The primary controlling feature of incentive systems is the inducement for practicing behavior consistent with performance objectives (Kerr 1988). Some work has found that reward system design and usage helps explain inter-organizational differences in successful change implementation (e.g., Agarwal and Singh 1998). Accountability is a critical element of incentive and reward systems (Bourdon 1982). Individuals are said to be accountable when their performance is monitored and when there are consequences (tangible or intangible) associated with the evaluation (Siegel-Jacobs and Yates 1996). Degree of accountability appears to affect decision-making and judgment. In particular, high levels of accountability appear to encourage more information gathering and examination and to lessen the possibilities of opportunistic behavior (Fandt and Ferris 1990; Hattrup and Ford 1995), and may be particularly important in motivating performance in situations of high interdependent behavior (Fandt 1991). Monitoring and control. Monitoring has long been considered a core activity of managers (e.g., Newman 1940).
Managers commonly employ diagnostic control systems (Anthony 1965) when monitoring planned change. In diagnostic control systems, managers gather information about the initiative of interest, assess the current state of performance against goals or objectives, and act on signicant differences between actual and desired performance (i.e., the performance gap) to achieve better results. As such, diagnostic controls help managers keep things on track (Merchant 1985, 1). The effectiveness of diagnostic control systems is reduced when comparative performance standards are imprecise or do not exist, or when output or behavior cannot be accurately measured (Lawler and Rhode 1976; Otley and Berry 1980). Despite its limitations, diagnostic control is thought to be central to the implementation of intended change, particularly those large in scale (Simons 1995).
Degree of accountability appears to affect decision-making and judgment.
Implementation success. Outcomes or results of a change initiative are frequently treated as a multidimensional variable. To assess the effectiveness of implementation, Tushman and OReilly (1997) suggested evaluating the extent to which the organization actually reached the intended future state, how well the organization functioned in its new state, and the cost of change to both organization and individual. Nadler and Tushmans (1980) congruence model, Tichys (1983) TPC framework, and the Burke-Litwin (1992) model all connect implementation success to both organizational performance and the effect or inuence on the individual. Miller (1997) proposed three dimensions that captured the degree of implementation success associated with a planned change: completion, achievement, and acceptability. Completion was the degree to which intended actions were implemented as planned. Achievement was the degree to which implemented actions were performed as intended. Acceptability was the degree to which the method of implementation and outcomes were satisfactory to those involved in, or affected by, the implementation. A well-rounded measure of implementation success, then, should assess change achievement at both the organizational and individual levels as well as dimensions that capture the notions of completion, achievement, and acceptability. Conguration Alternatives Using these factors as building blocks, we proceed to congure three alternative representations of the planned change process that are plausible expressions of existing theory. These congurations provide working models that can be subjected to empirical assessment.Diagrams of the three congurations appear in Figure 1. Direct effects model (M1). The most straightforward conguration of our ve building block factors involves simply linking each of the four variables of change process
61
Ford and Greer
FIGURE 1 Alternative Change Model Congurations
A. Direct Effect Model (M1)
AP
AP = Action Planning SD = Skill Development & Delivery I = Incentives M = Monitoring & Control IS = Implementation Success CP = Change Process (second order)
potential effects of several change process factors on implementation outcomes. M1 also expresses a non-sequential arrangement of the process variables. This specication supports the incremental, non-linear perspective of change process proposed by some theorists (e.g., Lindblom 1959; Quinn 1980). Congured in this fashion, change process factors such as action planning and skill development and delivery proceed mostly in parallel rather than in sequence to inuence implementation success. This model reects the following hypothesis: H1: Change process factors (action planning, skill development and delivery, incentives and monitoring and control) are positively related to implementation success. Second order change process model (M2). An alternative perspective views each change process variable as reecting a common, higher order change process construct (Figure 1b). Garvin (1998) viewed change processes as sequences of behaviors or events that altered the scale, character, or identity of the organization. From this perspective, a change process is more than just a collection of independent variables. Rather, the variables covary in a systematic way to reect the higher order construct. In this conguration, the gestalt effect of the variables is proposed as a more powerful way of predicting implementation success. This conguration emphasizes the overall strength of the organizations change process. Inside this process, the variables interact dynamically. M2s conguration de-emphasizes individual variables and stresses the organizations overall change process. This model reects the plausible notion that the process for achieving change may differ between organizations. Some organizations, for example, may realize successful change largely through exceptional planning while others rely heavily on effective reward systems. Organizations may differ in their prole of enacted change process variables while the relative strength of their overall change processes may be similar. These observations reect the following hypothesis: H2: Each change process factor (action planning, skill development and delivery, incentives and monitoring and control) reects a higher order change process construct that is positively related to implementation success. Sequential Model (M3). Implied in many models of planned change is a sequential progression that begins with planning activities and moves through variables that facilitate the execution of plans in order to realize effective change (e.g., Andrews 1971; Lewin 1951; Kotter 1996; Tichy 1983). While intuitively appealing, the notion that some actions must be done before others when implementing change has received surprisingly little research attention (Pettigrew, Woodman, and Cameron 2001). This investigations four change process factors can be categorized into
SD
IS
I M
B. Second Order Change Process (M2)
AP
SD CP I M
C. Squential Change Process (M3)
I
IS
AP
SD
IS
to the implementation success variable (Figure 1a). This conguration resembles a multiple regression model in which several independent variables are hypothesized to have a direct relationship on a single dependent variable. Several studies have employed this approach to examine the relationship between single change process variables and performance. Perhaps no change process variable has been studied in this fashion more so than planning, particularly in the context of its relationship to large-scale change achievement (e.g., Pearce, Robbins, and Robinson 1987). Metastudies of the confusing, often contradictory results owing from the stream of planning-performance studies have suggested that the models used to test hypothetical relationships require more accurate specication (e.g., Miller and Cardinal 1994). The M1 model specied here reects the 62
Ford and Greer
the three general stages of planned change process (Kanter, Stein, and Jick 1992). Action planning is a stage one activity that helps dislodge the organization from old patterns of behavior. Skill development is a stage two activity which serves to move the organization to new patterns of action. Incentives, and monitoring are stage three activities that govern behavior and help the organization institutionalize new patterns of action. Skill development, incentives, and monitoring can also be viewed as execution variables. These factors should be directly linked to implementation success since they make change happen by altering behavioral patterns in the organization. Levels of these factors should be related to action planning, since the plans provide the objectives and marching orders that must be operationalized. Moreover, since monitoring and incentives are mechanisms for governing behavior (Eisenhardt 1989; Fama1980), these two factors should also inuence skill development and delivery due to their institutionalizing character (Figure 1C). An interesting feature of this model is the mediating effect of three execution variables between action planning and implementation success. This planning execution outcomes sequence reects a common conceptualization of how intended organizational change occurs (e.g., Andrews 1971; Tichy 1983; Van de Ven and Poole 1995; Thompson and Strickland 1998) that is worthy of empirical testing in lieu of the rival view that such sequential order rarely occurs or is ill-advised (e.g., Mintzberg and Waters 1985). We should also note that the inclusion of execution variables in M3 highlights the role of implementation as a bridge between planning and performancea role thought by some to have been largely unaccounted for in the planningperformance studies (e.g., Smith and Kofron 1996). These observations reect the following hypotheses: H3a: Action planning is positively related to change process factors (skill development, incentives, monitoring). H3b: Change process factors (skill development, incentives, monitoring) are positively related to implementation success. H3c: Change process factors (incentives and monitoring) are positively related to skill development.
their organizations employed various activities during the implementation of a particular change in which they were involved. A complete description of the assessment process and the full questionnaire can be found in Center for Quality of Management (2001). We secured 107 useable questionnaires from individuals representing forty-three organizations. The primary unit of analysis in this study was an individuals assessment of the organizations change management processes in light of a specic planned change (individuals were asked to record this reference change in the questionnaire). During the data collection, individuals from the same parent organization often identied different initiatives to serve as their reference change, hence multiple respondent issues were not deemed an overly signicant concern. Indeed, the standard deviation between respondents in the full sample was found equal to or slightly higher than standard deviations between respondents in assorted sub-samples where respondent were restricted to one per organization. Sixtyfour percent of the respondents worked for service organizations and fty-six percent worked for manufacturing organizations. About two thirds of the respondents were from private, for-prot enterprises; others were about equally split between public, for-prot and public sector/government agencies. Ninety-two percent of the respondents were from organizations of more than 100 employees; 25 percent of respondents were from organizations of greater than 1000 employees. Over 90 percent of respondents were at least middle-level managers; more than half were upper-level managers. Questions designed to reveal stage and impact of planned change indicated that about 45 percent of the changes were estimated to be at least 50 percent completed at the time of the evaluation. Once implemented, over half of the planned changes were forecast to impact at least 40 percent of the organizations employees, suggesting that the majority of changes evaluated in this study were strategic, rather than incremental, in nature (see Nadler and Tushman 1989).
Measurement
Method
Sample data for this study were obtained from participants in change management seminars sponsored by the Center for Quality of Management. The Center for Quality of Management is an international consortium of over one hundred organizations focused on improving performance through the development and application of structured managerial processes. During the seminar, participants completed a questionnaire to assess the extent to which
Eleven items, those meant to reect the ve latent variables of our change model, were utilized from the questionnaire (Table 1). With 107 samples and eleven indicators, our ratio of samples to indicators was nearly 10:1, comfortably above the ve-to-one level often specied in multivariate studies (Hair, Anderson, Tatham, and Black 1998). The four independent latent variables of change process, action planning, skill development and delivery, incentives, and monitoring were each reected by two items (Table 1). As indicated by the associated alphas, each scale exhibited acceptable reliability. Response to each item consisted of ve choices organized on a Likert scale meant to reect the extent to which a formal system existed and was effectively implemented. A 1 represented little or no formal sysMid-American Journal of Business, Vol. 20, No. 2
63
Ford and Greer
TABLE 1 Indicators Used for Each Scale and Associated Reliabilities

Cronbachs Alpha Action Planninga AP1 Was an action plan developed for making the change? AP2 Was a timeline for successful completion established? Skill Development and Deliverya SD1 Did the organization develop necessary skills and capabilities through training, mentoring, outside acquisition or other means? SD2 Did the organization make sure that needed skills and capabilities were in place in time to complete the changes? Incentivesa I1 Were employees rewarded for working to support the change effort? I2 Were organization leaders held accountable for their behavior related to the change? Monitoringa M1 Was information effectively used to enable corrective action when necessary? M2 How effective were the actions taken to correct the progress of the change? .71
.76
.66
broadly employed in empirical studies of organizations (Nahm et al 2004; Ward and Duray 2000; King and Tao 2000). One method for evaluating whether response bias impairs the unidimensionality of measured variables is conrmatory factor analysis (Gerbing and Anderson 1988). A conrmatory factor analysis of the ve latent variable, 11-indicator measurement model representing the four independent change process variables of action planning, skill development and delivery, incentives, and monitoring and control, and the single dependent implementation success variable was conducted using LISREL 8 (Joreskog and Sorbom 2001). Signicant path coefcients (t values of 5.9 or higher) between each of the ve latent factors and their corresponding items resulted. Goodness of t statistics suggested acceptable model t1 (x2 = 42.19; df = 34; p = .158; RMSEA = .048; GFI = .93; AGFI = .87; NNFI = .96). Results from the conrmatory factor analysis suggested a measurement model with acceptable convergent and discriminant validity.
.74
Results
Implementation Successb .82 CA1 Did the change have a positive impact on business results? CA2 To what extent has the change resulted in expected behaviors? CA3 Overall, how satised were you with the changes? Response scale consisted of ve behaviorally anchored choices designed to reect the extent to which a formal system existed and was effectively implemented. A 1 represented an informal, ineffective system in place with few results; a 5 represented a formal effective system. b Response scale consisted of ve behaviorally anchored choices designed to reect the effectiveness of results achieved. A 1 represented little or no results to speak of; a 5 represented highly effective results.
a
tem in place with few results; a 5 represented a formal, effective system. Each response choice was behaviorally anchored to reduce the response scale drift that can confuse the detection of actual behavior changes when using questionnaires to measure change (Lindell and Drexle 1979). Implementation success was treated as a single dependent latent variable represented by three self-rated measures intended to reect the completion, achievement, and acceptability dimensions proposed by Miller (1997) (see Table 1). Responses consisted of ve behaviorally anchored choices meant to reect the effectiveness of results achieved. A 1 represented little or no results to speak of; a 5 represented highly effective results. Descriptive statistics and correlations for the eleven indicators used in this study appear in Table 2. Self-reported measures of performance are commonly noted as concerns due to the potential for common methods variance. However, self-reported measures have been
64
The three hypothetical change process model congurations were evaluated using the structural equation modeling methods of LISREL 8 (Joreskog and Sorbom 2001). Figure 2 includes the coefcients obtained from analysis of the direct effects model (M1). Only one of the path coefcients, the relationship between monitoring and control and implementation success, was found highly signicant (p < .001). The path coefcients between skill development and delivery and implementation success, and between incentives and implementation success, were found marginally signicant (p < .10). The path between action planning and implementation success was not signicant (t = 1.41). The squared multiple correlation for the implementation success latent variable was .70. Goodness of t statistics implied that the model t the data well. The chi-square was not signicant (x2 = 42.2; df = 34; p = .158). Additional indicators (RMSEA = .048; GFI = .93; AGFI = .87; NNFI = .96) met or exceedFIGURE 2 Analysis of Direct Effects Model of Change Process (M1)
x2 = 42.2, df = 34, p = .158 RMSEA = .048 GFI = .93 AGFI = .87 NNFI = .96 IS
p < .10 * p < .05 **p < .01 *** p < .001
AP .15 SD
.23 .24
I M
***.45
Ford and Greer
TABLE 2 Descriptive Statistics and Bivariate Correlations of Model Indicators
Mean AP1 AP2 SD1 SD2 I1 I2 M1 M2 IS1 IS2 IS3 * p < .05 ** p < .01 *** p < .001 Two tailed test 2.73 2.46 2.77 2.66 1.75 1.82 2.42 2.69 2.15 2.35 2.09
SDev 1.036 1.165 1.146 1.064 0.850 1.099 1.160 0.834 1.170 0.854 1.129
AP1 **.56 **.28 **.38 .18 .05 **.24 ***.28 **.25 ***.41 **.33
AP2
SD1
SD2
I1
I2
M1
M2
IS1
IS2
**.22 *.21 .19 .05 *.21 .20 *.23 **.30 *.26 **.61 *.24 .03 **.48 **.33 **.47 **.34 **.45 *.21 .12 **.37 *.27 **.33 **.28 **.44 **.50 **.35 **.29 **.40 **.29 **.29 .21 **.37 **.32 **.29 *.26 **.61 **.49 **.40 **.56 **.46 **.42 **.50 **.62 **.63 **.57
ed benchmarks indicative of reasonable t. These ndings suggest that Hypothesis 1 as stated should be rejected, since three of the four levers were found to have marginal or insignicant relationships to implementation success. Figure 3 includes the coefcients from analysis of the second order change process model (M2). All path coefcients were found highly signicant (p < .001). The strong path coefcient between the second order change process construct and implementation success supports the proposed relationship between this higher order change process variable and change achievement. The squared multiple correlation for the implementation success latent variable was .86. The t of this model was incrementally better than the t of M1. The chi-square statistic remained non-signicant (x2 = 46.2; df = 39; p = .201), and other indicators approached or exceeded benchmarks of reasonable t (RMSEA = .042;
FIGURE 3 Analysis of Second Order Model of Change Process (M2)
x2 = 46.2, df = 39, p = .201 RMSEA = .042 GFI = .93 AGFI = .88 NNFI = .97 ***.74
p < .10
AP
***.43
SD
***.59 ***.46
CP
IS
I M
***.69
* p < .05 **p < .01 *** p < .001
GFI = .93; AGFI = .88; NNFI = .97). These ndings suggest that Hypothesis 2 should not be rejected. Figure 4 includes the coefcients obtained from analysis of the sequential change process model (M3). While many of the path coefcients were found signicant, note that the strongest relationships surrounded the monitoring and control variable. The models sole insignicant path was between incentives and skill development, suggesting that incentives had little direct inuence on skill development and delivery, which provided only partial support for Hypothesis 3c. The squared multiple correlation for the implementation success latent variable was .68. While the t of this model was weaker than the t of the previous two models, the t statistics remained at acceptable levels (x2 = 55.5; df = 36; p = .02; RMSEA = .071; GFI = .91; AGFI = .84; NNFI = .92).1 These ndings suggest that Hypotheses 3a and 3b should not be rejected. To summarize, analysis of M1s conguration found monitoring and control as the most signicant change process lever linked to implementation success. Hypothesis 1 should be rejected, since action planning, skill development and delivery, and incentives were also proposed as directly related to change achievement. Analysis of M2s conguration found highly signicant paths (p < .001) emanating from the second order change process construct to the other model variables, including implementation success. The strong path coefcients and measured t support Hypothesis 2s notion the gestalt effects of individual levers as part of a higher level change process construct. The strength of the path coefcients and measures of overall t also provided reasonable support for the sequential ordering of change process variables (M3) as proposed by Hypotheses 3a, 3b, and 3c. Of particular note was the strong relationship between the monitoring and control factor and other variables 65
Ford and Greer
FIGURE 4 Analysis of Path Model of Sequential Change Process (M3)

x2 = 55.5, df = 36, p = .020 RMSEA = .071 GFI = .91 AGFI = .84 NNFI = .92
I
*.31
.28
*.29 -.01
AP
SD
***.50 ***.46
*.27
IS
***.49
p < .10 * p < .05 **p < .01 *** p < .001
of the model, which suggests the importance of this variable on the achievement of planned change.
Viewing the change process using a perspective similar to the direct effects M1 conguration in Figure 1 appears nave. Strength of the results from the analysis of M2 and M3 suggests a more dynamic perspective of the planned change process. One such perspective is that of a highlevel change process construct which captures patterns of covariation among the individual change process variables. It is consistent with the non-linear path through which many changes are realized (e.g., Lindblom 1959; Quinn 1980). Emphasizing the higher order change process construct rather than the individual levers supports a view that different organizations might emphasize different levers at their disposal for the implementation of change. Such a perspective is intuitively appealing since it emphasizes the uniqueness by which each organization might approach the implementation problem. Our ndings also suggest the plausibility of modeling some sequential organization among change process variables. Many researchers have suggested that the process of change is sequential to some degree, and that, when implementing change, it is more important to alter some elements of the organization before others (e.g., Hinings and Greenwood 1988; Gersick 1994). Our ndings suggest that this is a reasonable view from a measurement perspective, which should motivate further inquiry into causal order among change process variables. Of the four independent change process variables considered in this investigation, monitoring and control appeared 66
Discussion
to have the strongest effect on implementation success. The path coefcients associated with the monitoring and control variable were relatively strong in each of the three models examined (see Figures 1-3). Monitorings salience to change achievement may relate to the dynamic, revisionist nature of planned change. Most planned changes, particular those large in scale, require midstream corrections to the initial course of action (Mintzberg and Waters 1985), which may necessitate formal monitoring of implementation progress. This study has some limitations. Our sample was conned to respondents from organizations that were members of the same industrial organization. Moreover, the sample size was relatively small in comparison to other multivariate studies, and included respondents from the same organization. While the resulting demographics of the sample appeared reasonable and multiple respondent inuences were deemed minimal, a larger, broader sample would be desirable in future studies. By design, the change variables selected for this study were limited to a few widely accepted factors in order to explore some fundamental empirical questions. Of course, other factors have been proposed to impact the process of change, such as climate and culture (Burke and Litwin 1992), previous decision history (Nadler and Tushman 1980), politics (Tichy 1983), and communication (Kotter 1995). Entering additional factors of change process would make for a more comprehensive analysis. In addition, the two- and three-item measurement scales were smaller than those often employed in structural equation modeling studies. Future research could explore larger measurement scales to round out the content validity of the model. Finally, researchers have noted concerns with self-rated measures of change, based largely on the argument that a raters basis for comparison shifts as the organization itself changes (e.g., Zmud and Armenakis 1978). While objective measures of organizational change are certainly desirable, nding them has been problematic for both researchers (Cameron 1980; Lewin and Minton 1986) and practitioners (Troy 1994). We should note that despite such concerns, self-rated measures have been effectively employed in a number of insightful implementation studies (e.g., Nutt 1986; Miller 1997; Nahm,Vonderembse and Koufteros 2004). This study suggests the value of survey-based empirical research for studying organizational change. Pettigrew, Woodman, and Cameron (2001) identied issues related to temporality, sequencing, and linkage to organizational outcomes among the challenges facing researchers of organizational change. Although researchers often suggest only qualitative or case based methods for gathering change process knowledge, survey-based empirical research can help researchers pursue such issues. For example, periodically gathering questionnaire-based data over the life of an implemented change could provide insight into when organizations employ particular change process factors, the degree to which such factors were employed, and how outcomes responded to the various process adjustments.
Ford and Greer
As noted previously, the validity of M2s conguration raises the possibility that organizations possess unique change process proles for implementing planned change. The prole, reective of the degree to which various process variables are enacted during implementation, might relate to the organizations particular set of skills or competences (e.g., Barney 1991). Organizations that possess strong communication skills, for instance, might emphasize change process factors with high communication content to a greater extent than less uent organizations. Empirical designs could investigate the existence of such change process proles and the extent to which they may be linked to an organizations underlying resources. In our study, monitoring and control were consistently found to be related to implementation success. Our study has practical implications for managers accountable for successfully implementing planned change. Findings from our evaluation of M2 suggest the possibility of developing a unique change process for each organization. Instead of subscribing to one particular set of change process factors, it appears plausible that an organization might be able to develop their own change style or combination of process factorsperhaps based on particular organizational skills or strengths. For example, an organization with poor planning skills might still realize implementation success if it can compensate with effective skill development and delivery during the change process. In addition, our evaluation of M3 suggests some sequential character to the change process, which supports the notion that timing or pace may be an important consideration when implementing change (Gersick 1994). Do some change process factors matter more than others? In our study, monitoring and control were consistently found to be related to implementation success. Since modications to an initial course of action are highly probable (Mintzberg and Waters 1985), diagnostic control systems may be essential for managers to detect performance gaps that impair implementation success and require corrective action. Effective monitoring and control requires organizational skills in objective setting, in information retrieval and analysis, and in selecting the appropriate corrective action if a signicant deviation from plan is detected (Simons 1995). Many of these skills are similar to factors thought to embody learning organizations (Nevis, DiBella, and Gould 1995). In other words, an organizations effectiveness in diagnostic monitoring and control may reect general capacity for organizational learning and change management (Kloot 1997). Managers who are accountable for change outcomes might benet from establishing monitoring and control systems that permit tracking of implementation progress and effective intervention when necessary.
Finally, we should note that, while a number of researchers have also observed the empirical importance of monitoring and control in achieving change (e.g., Charan and Colvin 1999; Kotter and Schleisinger 1979), many elements of modern organization design may not be conducive to monitoring. Managers have been busy shedding bureaucracy, decentralizing decision-making, and establishing more workplace autonomy to help the organization move faster and become more innovative (Burns and Stalker 1961; Ouchi 1980). Although such practices might help get change going, lack of formal control structure might impair effective execution of the plan. Such a premise is consistent with the Were great starters, but terrible nishers assessment we often hear from managers characterizing the change processes in their organizations. Many organizations may be reaching or exceeding advisable limits for decentralized control (Bungay and Goold 1991). Further inquiry into the role of monitoring in the process of change is prudent.
Conclusions
Given the dynamic work environments that exist in most organizations, understanding planned change and its components for success is a necessary skill for managers involved in implementing short and long-term strategic objectives. Indeed, the growing use of the term execution in the lexicon of management (e.g., Bossidy 2003) suggests that the value placed on knowing how to manage change is increasing. Knowledge about change models and the factors that compose them can only benet managers who must pull the proper levers that lead to successful implementation. The ndings from this study support congurations that reected dynamic change process conceptualizations. The dynamic change processes were found to possess favorable measurement properties when compared to a direct effects model. Of the change process variables considered in this investigation, monitoring and control demonstrated the strongest relationship to implementation success. These ndings support a dynamic, perhaps sequential perspective of change and its implementationa perspective that should benet from further empirical investigation.
Notes
1. We utilize several commonly reported goodness of t indicators and the thresholds suggested by Hair et al. (1998) as desirable. x2 is the chi-square statistic (a non signicant p-value of at least p > .01 is desirable). RMSEA is root mean square error of approximation (< .08). GFI is goodness of t index (no consensus threshold but .90 often viewed as minimum acceptable value). AGFI is adjusted goodness of t index (> .85). NNFI is non-normed t index, also known as the Tucker-Lewis index (> .90).
67
References
Adams, J.A. 1967. Human memory. New York: McGraw-Hill. Agarwal, N. and P. Singh. 1998. Organizational rewards for a changing workplace: An Examination of theory and practice. International Journal of Technology Management 16: 225-238. Alchian, A.A. 1950. Uncertainty, evolution, and economic theory. Journal of Political Economy 58: 211-221. Andrews, K.R. 1971. The concept of corporate strategy. Homewood: Dow Jones-Irwin, Inc. Ansoff, H.I. 1965. Corporate strategy: An analytical approach to business policy forgrowth and expansion. New York: McGraw Hill Book Company. Anthony, R.N. 1965. Planning and control systems: A framework for analysis, Boston: Division of Research, Graduate School of Business Administration, Harvard University. Baldwin, T.T. and R.J. Magjuka. 1997. Training as an organizational Episode: Pretraining inuences on trainee motivation. In J.K. Ford, Ed., Improving Training Effectiveness in Work Organizations. 99-127. Mahwah: Erlbaum Barnard, C.I. 1938. The functions of the executive. Cambridge: Harvard University Press. Barney, J.B. 1991. Firm resources and sustained competitive advantage. Journal of Management 17: 99-120. Bossidy, L. and R. Charan. 2002. Execution: The discipline of getting things done. NewYork: Crown Business. Bourdon, R. 1982. Measuring and tracking performance management for accountability. Journal of Organizational Behavior Management 4(3-4):101-113. Bungay, S. and M. Goold. 1991. Creating a strategic control system. Long Range Planning 24(3):32-39. Burke, W.W. and G.H. Litwin. 1992. A causal model of organizational performance and change. Journal of Management 18: 523-545. Burke, W.W. 1995. Diagnostic models for organization development. In A. Howard & Associates Eds., Diagnosis for organizational change 53-84. New York: Guilford Press. Burns, T. and G.M. Stalker. 1961. The management of innovation. London: Tavistock Publications. Cameron, K. 1980. Critical questions in assessing organizational effectiveness. Organizational Dynamics 9: 66-80. Center for Quality of Management. 2001. Mobilizing change using the 7 infrastructures. Cambridge: Center for Quality of Management. Charan, R. and G. Colvin. 1999. Why CEOs fail. Fortune 139(2):68-78. Cottrill, M. 1997. Give your work teams time and training. Academy of Management Executive 11(3): 87-89. Cummings, L.L. and D.P. Schwab. 1973. Performance in organizations: Determinants and appraisal. Glenview: Scott, Foresman. Eisenhardt, K.M. 1989. Agency theory: An assessment and review. Academy of Management Review 14: 57-74. Fama, E. 1980. Agency problems and the theory of the rm. Journal of Political Economy 88: 288-307. Fandt, P.M. and G.R. Ferris. 1990. The management of information and impressions: When employees behave opportunistically. Organizational Behavior and Human Decision Processes 45: 140-158. Fandt, P.M. 1991. The relationship of accountability and interdependent behavior to enhancing team consequences. Group & Organization Studies 16: 300-313. Garvin, D.A. 1998. The processes of organization and management. Sloan Management Review 39(4):33-50.
Gerbing, D.W. and J.C. Anderson. 1988. An updated paradigm for scale development incorporating unidimensionality and its assessment. Journal of Marketing Research 25:186-192. Gersick, C.J.G. 1994. Pacing strategic change: The case of the new venture. Academy of Management Journal 37: 37-45. Gioia, D.A. and K. Chittipeddi 1991. Sensemaking and sensegiving in strategic change initiation. Strategic Management Journal 12: 433-449. Goodman, P.S. and J.W. Dean Jr. 1982. Creating long-term organizational change. In P.S. Goodman, Ed., Change in Organizations 226-279. San Francisco: Jossey-Bass Hair, J.F., R.E. Anderson, R.L. Tatham, and W.C. Black. 1998. Multivariate data analysis, 5th Ed. New York: Macmillan Publishing Co. Hannan, M.T. and J. Freeman. 1977. The population ecology of organizations. American Journal of Sociology 82: 929-964. Hattrup, K. and J.K Ford. 1995. The roles of information characteristics and accountability in moderating stereotype-driven processes during social decision making. Organizational Behavior and Human Decision Processes 63:73-86. Hinings, C.R. and R. Greenwood. 1988. The dynamics of strategic change. Oxford:Blackwell. Hofer, C.W. and D. Schendel. 1978. Strategy formulation: Analytic concepts. St Paul: West Publishing Company. Huy, Q.N. 2001. Time, temporal capability, and planned change. Academy of Management Review 26: 601-623. Jensen, M. and M. Meckling. 1976. Theory of the rm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3: 305-360. Joreskog, K.G. and D. Sorbom. 2001. LISREL 8: Users reference guide. Lincolnwood: Scientic Software International, Inc. Kanter, R.M., B.A. Stein and T.D. Jick. 1992. The challenge of organizational change. New York: Free Press. Kerr, J.L. and E.F. Jackofsky. 1989. Aligning managers with strategies: Management development versus selection. Strategic Management Journal 10:157-170. Kerr, J.L. 1988. Strategic control through performance appraisal and rewards. Human Resource Planning 11: 215-223. King, W. R., & T.S.H Teo. 2000. Assessing the impact of proactive versus modes of strategic information systems planning. OMEGA: International Journal of Management Science 28:667-679. Kloot, L. 1997. Organizational learning and management control systems: Responding to environmental change. Management Accounting Research 8:47-73. Kotter, J.P. and L.A. Schleisinger. 1979. Choosing strategies for change. Harvard Business Review 57(2): 106-114. Kotter, J.P. 1995. Leading change: Why transformation efforts fail. Harvard Business Review 73(2): 59-67. Kotter, J.P. 1996. Leading change. Boston: Harvard Business School Press. Lawler, E.E. III and J.G. Rhode. 1976. Information and control in organizations. Santa Monica: Goodyear. Levy, A. 1986. Second-order planned change: Denition and conceptualization. Organizational Dynamics 15(1): 5-20. Lewin, K. 1947. Frontiers in group dynamics. Human Relations 1:5-41. Lewin, K. 1951. Field theory in social science. New York: Harper. Likert, R. 1961. New patterns of management. New York: McGraw-Hill. Lindblom, C. 1959. The science of muddling through. Public Administration Review 19(2): 79-88.
68
Lindell, M.K. and J.A. Drexler Jr. 1979. Issues in using survey methods for measuring organizational change. Academy of Management Review 4:13-28. Lippit, R., J. Watson, and B. Westley. 1958. The dynamics of planned change. New York: Harcourt Bruce. Merchant, K.A. 1985. Control in business organizations. Marsheld: Pitman. Miller, C.C. and L. Cardinal 1994. Strategic planning and performance: A synthesis of more than two decades of research. Academy of Management Journal 37:1649-1665. Miller, S. 1997. Implementing strategic decisions: Four key success factors. Organization Studies 18: 577-602. Mintzberg, H. and J.A. Waters. 1985. Of strategies, both deliberate and emergent. Strategic Management Journal 6: 257-272. Nadler, D.A. and M. Tushman. 1980. A model for diagnosing organizational behavior: Applying the congruence perspective. Organizational Dynamics 9(2):35-51. Nadler, D.A. and M. Tushman. 1989. Organizational frame bending: Principles for managing reorientation. Academy of Management Executive 3(3):194-204. Nahm, A, M Vonderembse and X Koufteros. 2004. The impact of organizational culture on time-based manufacturing and Performance. Decision Sciences 35(4):579-607. Nevis, E.C., A.J. DiBella, and J.M. Gould. 1995, Understanding organizations as learning systems. Sloan Management Review 36(2): 73-85. Newman, W.H. 1940. Business policies and management. Cincinnati: Southwest Publishing Co. Nutt, P.C. 1986. Tactics of implementation. Academy of Management Journal 29:230-261. OReilly, C and B. Weitz. 1980. Managing marginal employees: The use of warnings and dismissals. Administrative Science Quarterly 25: 467-484. Otley, D.T. and A.J. Berry. 1980. Control, organization, and accounting. Accounting, Organizations and Society 5: 231-246. Ouchi, W.G. 1980. Markets, bureaucracies, and clans. Administrative Science Quarterly 25:125-160. Pearce, J., D. Robbins and R. Robinson. 1987. The impact of grand strategy and planning formality on nancial performance. Strategic Management Journal 8:125-134. Pettigrew, A.M., R.W. Woodman and K.S. Cameron. 2001. Studying organizational change and development: Challenges for future research. Academy of Management Journal 44:697-713. Porras, J.I. and S.J. Hoffer 1986. Common behavior changes in successful organization development. Journal of Applied Behavioral Science 22:477-494. Quinn, J. 1980. Strategies for change: Logical incrementalism. Homewood: Irwin. Robertson, P.J., D.R. Roberts and J.I. Porras. 1993. Dynamics of planned organizational change: Assessing empirical support for a theoretical model. Academy of Management Journal 36: 619-634. Romanelli, E. and M.L. Tushman. 1994. Organizational transformation as punctuated equilibrium: An empirical test. Academy of Management Journal 37:1141-1166. Siegel-Jacobs, K. and J.F. Yates. 1996. Effects of procedural and outcome accountability on judgment quality. Organizational Behavior and Human Decision Processes 65:1-17. Simon, H.A. 1947. Administrative behavior. 2nd ed, New York: The Macmillan Company. Simons, R. 1995. Levers of control. Boston: Harvard Business School Press.
Smith, K.A. and E.A. Kofron. 1996. Toward a research agenda on top management teams and strategy implementation. IBAR Irish Business and Administrative Research 17(1): 135-152. Tannenbaum, R. 1971. Organizational change has to come through individual change. Innovation 23: 36-43. Terpstra, D.E. and E.J. Rozell. 1993. The relationship between stafng practices to organizational level measures of performance. Personnel Psychology 46(1): 27-48. Thompson, A.A., Jr. and A.J. Strickland III 1998. Strategic management: Concepts and cases. 10th ed., Boston: Irwin McGraw-Hill. Tichy, N.M. 1983. Managing strategic change. New York: Wiley. Troy, K. 1994. Change management: An overview of current initiatives. New York: The Conference Board. Tushman, M.L. and C.A. OReilly. 1997. Winning through innovation: A practical guide to leading organizational change and renewal. Boston: Harvard Business School Press. Van de Ven, A.H. and M.S. Poole. 1995. Explaining development and change in organizations. Academy of Management Review 20:510-540. Venkatramin, N. 1990. The concept of t in strategy research: Toward verbal and statistical correspondence. Academy of Management Review 14:423-444. Ward, P. T., and R. Duray. 2000. Manufacturing strategy in context: Environment, competitve strategy and manufacturing strategy. Journal of Operations Management 18:123-138. Weick, K. and R. Quinn. 1999. Organizational change and development. Annual Review of Psychology 50:361-383. Weisbord, M.R. 1976. Organizational Diagnosis: Six Places to Look for Trouble With or Without a Theory. Group and Organization Studies 1: 430-447. Werr, A. 1995. Approaches, methods and tools of changeA literature survey and bibliography. Economic and Industrial Democracy 6:607-651. Zmud, R.W. and A.A. Armenakis. 1978. Understanding the Measurement of Change. Academy of Management Review 3:661-669.
About the Authors

Dr. Matthew W. Ford is an Assistant Professor of Management at the Northern Kentucky University College of Business. He holds a Ph.D. in Operations Management from University of Cincinnati. His research interests include quality management, entrepreneurship, and the management and control of change. fordmw@nku.edu Dr. Bertie M. Greer is an Assistant Professor of Management at Northern Kentucky University. She holds a Ph.D. in Business from Kent State University. Dr. Greers teaches courses in operation management, project management and total quality management. Dr. Greers research interests include change management, supply chain management, quality management, project management, and diversity. greerb@nku.edu
69
70
Cummings
BOOKSHELF
Final Accounting
By Barbara Ley Tofer Reviewer: Bill Cummings Northern Illinois University What do all the following organizations have in common: Baptist Foundation of Arizona, Global Crossing, Sunbeam, Waste Management, WorldCom, and Enron? All were clients of the Arthur Andersen (simply Andersen after May of 2001) accounting and auditing rm during the period of 1990 through 2002. In addition, all suffered major frauds, signicant, downward restatements of earnings, and sanctions by federal regulators. Much has been written over the past three years about the fall of Andersen, the once proud and highly respected giant in the accounting and auditing world. Final Accounting by Barbara Tofer may be the best account available of the events leading to the demise of Andersen as the auditor of many of the worlds best known companies. Ms. Tofer, as the partner-in-charge of Ethics and Responsible Business Practices consulting services for Andersen from 1995 to 2000, is in the unique position of being able to tell the story of Andersens demise from the inside and as a professional ethics expert. The author chronicles the history of the rm from its founding in 1913 by Northwestern University accounting professor Arthur Andersen through the early entry of the rm into the consulting world in the 1940s right up to the events leading to the breakup of the auditing and consulting practices in the early 2000s. This breakup she contends was the proximate cause of Andersens eventual fall from grace and lead to the indictment and conviction for obstruction of justice in the Enron case. A conviction, which ironically at the time of this writing, has been reversed by the U.S. Supreme Court. But the story of Andersens demise and fall is like a modern, tragic, morality play how the once most powerful and highly respected accounting rm in the world lost its moral compass. Then, born out of desperation to rebuild its revenue base after the departure of its consulting arm, Andersen Consulting, now known as Accenture, frantically pushed its partners to cross-sell services beyond the areas of accounting and auditing. 70
The major point of the book is to show how the Enron scandal alone did not bring down Andersen. Rather, commitment to audit excellence and to high ethical standards, which had been the hallmarks of Andersen for its entire history, began to erode over a decade before the actual demise of the rm. Portrayals in the media during 2002, when Andersens fall was unfolding, focused on Andersens failure to reveal the ongoing fraud in Enron and alleged acts of obstruction of justice which led to the indictment by the Justice Department of the entire rm. Almost everyday in the business press were new reports of longtime Andersen clients dumping the rm as their auditor and switching to another CPA rm. Indeed, the nal destruction of the rm as an auditor of public companies was due to the convergence of several factors which made Andersens survival impossible: the media frenzy in the spring of 2002 regarding Enron and Andersen, Andersens almost totally discredited reputation with further revelations regarding Enron and other audit failures, and the fact that most public corporations hold their annual shareholder meetings in the spring. Given this perfect storm of events, it is hard to imagine any audit committee chairman or chief nancial ofcer announcing at the annual meeting that Andersen had been retained as the outside auditor. Not one company stood by the rm. The ight of clients from Andersen in the spring of 2002 resembled a classic run on a bank. But the complete story of Andersens decline and fall is much longer and more involved. Tofers account weaves together the authors personal experiences at Andersen with the bigger picture of how the rms demise began and ultimately unfolded. In so doing, she identies some of the major lessons that accountants, educators, and the investing public should learn from the Andersen saga. The following are examples of several of major lessons raised by Tofer: Everyone Thinking the Andersen Way Did Not Serve the Firm Well in the Long Run One of the hallmarks of Andersen had always been the tendency to hire people right out of college and intensively train them in the Andersen rules of the game. This approach was originally conceived by Arthur Andersen, the founder and furthered by Leonard Spacek who followed as the rms managing partner. The idea was that no matter who you were or where you did business in the world, as a client you would receive the same quality of service and approach to work from people who were all trained in the same basic way. Andersens St. Charles, Illinois training facility in fact was unique in the industry. New employees and veterans alike from all over the world were brought to St. Charles to indoctrinate and refresh them in the Andersen approach to accounting, auditing, and consulting practices. The Andersen way included a dress code (in the 1950s and 1960s conservative dark suits, white shirts, and a hat when outside) and a mandate always be busy and to walk briskly. But even more important, Andersen traditions placed a premium
Cummings
on not upsetting clients and not questioning superiors. When the rm appeared to be losing its ethical halo in the 1990s, these latter two traditions made it extremely difcult for the rm to recognize the gravity of the situation or for anyone to initiate action to correct the mistakes of wrongful conduct. The rm rarely looked to outsiders to conduct training but rather used Andersen people almost exclusively. People who could not adopt and internalize the Andersen way moved out to clients or other accounting rms. All of this had the effect of creating a very consistent but insular Andersen workforce. In fact, Andersen people became known in the industry as Androids because of their unswerving loyalty to the rms way of doing everything. In the end the Android mindset also led to arrogance that the rms way was always the best and denial that the rm could ever do anything wrong. Up to the very end in national media interviews, Joseph Bernadino, the nal managing partner of the rm, denied that Andersens actions constituted wrongful conduct. Active employees and retirees alike were dumbfounded that the federal government could or would indict the rm essentially putting it out of business. The Justice Department position was quite different of course, to them, Andersen was a repeat offender. Having a series of audit failures in the 1990s and having signed a consent degree in 2001 to clean up its practices after the Waste Management debacle, Andersen had not carried through on its part of the bargain. The nal death knell in the court of public opinion was the announcement by World Com, another client, of a $9 billion downward restatement and the subsequent arrest of World Coms CEO and CFO for fraud. All that was left for Andersen to do was surrender its licenses to practice and handoff its clients to competitors. The Split between Auditing and Consulting The Beginning of the End for Andersen What had long been touted as Andersens great strength in the accounting industry, the size of its consulting practice, probably planted the seeds of the rms ultimate destruction. Tofer considers the rapid growth, the surpassing of audit revenue by that of consulting, and the eventual break away of Andersen Consulting to be several of the primary causes of Andersens auditors to lose their ethical compass and bring down the rm. Andersen had been an early entrant into the consulting eld and one of the rst to provide computer and systems consulting services to its clients. The business was so successful that the growth rate of consulting compared to auditing was signicantly greater by the 1980s. This ultimately led to a cultural divide which resembled the Grand Canyon as the consulting partners became convinced that it was their side of the business which was in essence subsidizing the audit side of the rm. The divide deepened and became bitter during the late 1980s and in an effort to keep the rm together, two autonomous business units were formed, Arthur Andersen (AA) for the audit side and
Andersen Consulting (AC) for the consulting business, both under the umbrella of Andersen Worldwide. But the One Firm concept had been effectively breached. The two sides of Andersen coexisted in the 1990s in a contentious fashion more resembling the Cold War than a collegial rm. Without ACs revenue, AA was now at or near the bottom of the six major accounting rms in size. In an effort to rebuild its revenue base, AA started its own consulting arm ostensibly for smaller, emerging business. Inevitably however, there was some direct competition for clients with AC and this fact was not trivial to ACs partners who saw the encroachment by AA as bad faith especially since AC still had to transfer 15 percent of its prots to AA. In addition to its foray into consulting, AA began to press its partners to aggressively market and even crosssell services to clients. Billing our brains out is the way Tofer described the strategy to rebuild the revenue base. Is this where Andersen, the one-time ethical leader of the eld, began to lose its way? One can only surmise that as revenue building became the preeminent goal of the rm, quality auditing, and upholding the staunch ethical standards of the rm became secondary. Certainly, the string of audit failures involving Andersen clients in the 1990s, points to something going terribly wrong. As the friction between AA and AC worsened, in late 1997, AC sued for divorce by voting to break from AA and ling for arbitration to determine how much AC would have to pay for its freedom. Andersen demanded almost $15 billion from AC for the break; the arbitrator ruled that AC would have to pay about $1 billion and cease to use the Andersen name which was trivial in that AC had already ceased using the name and would soon become Accenture. The nal loss of AC for such a low price was devastating and now given the leadership chaos which existed at the rm, the stage was set for Andersens nal slide into disgrace and dissolution. In this writers opinion, the best part of Final Accounting is the nal chapter entitled Other Peoples Money. In this chapter, Tofer states what she believes are the true lessons from Andersens demise and fall. For example, did Andersen self-destruct because of a few bad apples in the barrel or is the whole environment which spawned disasters like Enron and WorldCom rotten? Certainly, all frauds and audit failures are perpetrated by individuals but the point that Tofer is making here relates to the vulnerability of Andersens people to the intense pressures to which todays accountants and auditors are subject. The extreme importance that the capital markets place on corporations to meet expected earnings targets can translate into intense management pressure on the CPAs during the audit process. It is certainly not accidental that this decade has seen record numbers of earnings restatements and nancial reporting frauds by clients of all the major accounting rms. But was Andersen more vulnerable to pushing the envelope, given the contention with and eventual loss of AC, rapid management turnover at the top of Andersen, and the almost desperate clamor to build revenues of the rm? 71
Cummings
The failure of Andersen is a very powerful lesson to us all: accountants, business people, and educators alike. Business organizations must make ethical behavior a top priority and carry through on that priority from top to bottom. For the CPA profession, protecting the public interest must be the primary purpose of the audit practice. There is nothing inherently wrong with CPAs practicing as consultants but consulting cannot be the driving force in the rm relationship with the client. Arthur Leavitt, former chair of the SEC, was undoubtedly correct in his assessment that one rm cannot act as auditor and consultant to the same company without ethical conicts eventually surfacing and possibly compromising the audit. In the end, what the CPA has to offer is competence and integrity in protecting the investing publics interest. When that competence and integrity is irreparably damaged in the eyes of the public, that rm will not survive. For business and accounting professors, the case of Andersen and Enron has provided many rich classroom examples and chances for discussion. But the lesson is clear to us as well, students and faculty must be aware of the ethical dilemmas that we face in life and in the practice of our profession. Even an accounting rm, which many educators regarded in the highest sense for its professionalism and ethical standards, can lose its moral compass. We must also have strategies for how we will deal with these ethical situations when they confront us. Clearly when Andersen was confronted by immense ethical problems, it lacked the basic ability to resolve those conicts in a morally defensible way. Tofers Final Accounting is an excellent account of how a once highly regarded organization can lose its ethical sense of balance and fail as protector of the public interest. I highly recommend the book to anyone interested in business ethics or the Andersen/Enron case specically.
ACCESS QUALITY BUSINESS RESEARCH

The Mid-American Journal of Business is proud to be listed in:
Abstracting Ser vices H.W. Wilson (BPI) ABI/INFORM Global Omnile (H. W. Wilson Company) The Business Index(online) Ziff Business and Economics Research Directory Cabells Directory of Publishing Oppor tunities in Management and Marketing Emerald Reviews Publishers Consulting Corporation The Serials Directory--EBSCO Standard Periodical Directory Ulrichs International Periodicals Directory Working Press of the Nation PsycINFO, PsycLIT, ClinPYSC, and Psychological Abstracts American Overseas Book Company Gale Directory of Publications and Broadcast Media InfoTrac
Subscription Ser vices EBSCO Subscription Service Harrassowitz Booksellers & Subscription Agents McGregor Subscription Service Popular Subscription Service Readmore Subscription Service Reginald F. Fennell Subscription Service SWETS Blackwell Subscription Service
www.bsu.edu/MAJB
72

Vol 20 Num 2

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Vol 20 Num 2

Загружено:

Авторское право:

Доступные форматы

MID-AMERICAN JOURNAL OF BUSINESS

IBBEA Implementation and the Relative Protability of Small Banks

Modeling Internet Operations Using Initial Public Offerings

Building a Market-Oriented Organizational Environment: An Implementation Framework for Small Organizations

Implementing Planned Change: An Empirical Comparison of Theoretical Perspectives

Managing Our Way to Academic Decline

Ashok Gupta Editor-in-Chief

Mid-American Journal of Business, Vol. 20, No. 2

Business Colleges Should Practice What They Preach

Become more entrepreneurial

Live your strategic plan

Measure how well your college is performing

The Mid-American Journal of Business

Improve the skills of your workforce

Central Michigan University

Know your customer and markets

Northern Illinois University

The University of Toledo

Western Michigan University

Mid-American Journal of Business, Vol. 20, No. 2

Corporate Culture Denes a Company and its Future

Richard T. Farmer Chairman and Founder, Cintas Corporation

Why Culture Counts

Mid-American Journal of Business, Vol. 20, No. 2

Mid-American Journal of Business, Vol. 20, No. 2

Mid-American Journal of Business, Vol. 20, No. 2

MIDWEST DECISION SCIENCES INSTITUTE

2006 Annual Meeting

March 23-25, 2006

Call for Papers

CALL FOR PAPERS

Instructions for Contributors

BEST PAPER AWARD

Submission Deadline: January 17, 2006

Mid-American Journal of Business, Vol. 20, No. 2

An Analysis of Working Capital Management Results Across Industries

Filbeck and Krueger

Table 1 Industries Represented in CFOs Working Capital Management Surveys

Table 2 Working Capital Management Component Denitions and Averages

Cash Conversion Efciency (CCE)

(Cash ow from operations) / Sales

Days Working Capital (DWC)

Days Payables Outstanding

Accounts Payable / (Sales/365)

Mid-American Journal of Business, Vol. 20, No. 2

Filbeck and Krueger

Mid-American Journal of Business, Vol. 20, No. 2

Filbeck and Krueger

Mid-American Journal of Business, Vol. 20, No. 2

Filbeck and Krueger

Mid-American Journal of Business, Vol. 20, No. 2

Filbeck and Krueger

Mid-American Journal of Business, Vol. 20, No. 2

Filbeck and Krueger

Mid-American Journal of Business, Vol. 20, No. 2

Filbeck and Krueger

About the Authors

Mid-American Journal of Business, Vol. 20, No. 2

IBBEA Implementation and the Relative Protability of Small Banks

Mid-American Journal of Business, Vol. 20, No. 2

Nippani and Washer

Table 1 Number of Banks in Various Size Categories at Different Points in Time

Mid-American Journal of Business, Vol. 20, No. 2

0.162 0.303 0.230**