Вы находитесь на странице: 1из 11

CIS 9000 INSTRUCTOR: PROFESSOR M PALLEY APRIL 24TH, 2007

CASE ANALYSIS

ROSHAN ROBERT MARTA KAMIENIECKI PRIYADHARSINI K YAGESH RAWAT GLADYS ANNAN

I.

Executive Summary

The Food Services Company, SYSCO has decided to purchase Business Intelligence software, (BI) from the firm-Business Objects. This is a natural progression in SYSCOs continuous initiatives to streamline the issues of data redundancy and data integrity that has arisen due to the decentralized operational structure of the company. The IT initiatives in this regard have been: ERP development and deployment Data warehousing effort The logical progression in this initiative is to develop methodologies to leverage this data to provide superior monitoring and analytical capabilities. In analyzing the case we considered the dichotomy between the decentralized functionality of the operating companies of SYSCO and a centralized approach that the business intelligence architecture would facilitate. We also took cognizance of change management issues that could crop up as a result of this, which have not been addressed by SYSCO. The value of Business intelligence increases as the delivery of information is embedded in the processes and systems of the enterprise. In this regard, the well-crossthat-bridge-when-we-come-to-it stance taken by SYSCO can have a long term detrimental effect. We propose moderate changes in the road ahead. After a critical analysis of SYSCOs business model we concluded that effective predictive analysis would have to take into consideration the internal and external environment. In the internal environment valuable intelligence can be gained only by performing integrated analytics that cover cross value chain information across the following spectrum: Customer analytics. Relationship management Supply chain management. We propose a data to information on demand approach that adds the integrated analytics layer to the data platform. To facilitate this and to take into consideration the imminent change management issues as well as to implement predictive analysis in the customer analytic-relationship management-supply chain business model(with inputs from the external environment)-we propose a three tier BI implementation: Tier 1-Open adaptive investigative-For non recurring business questions Tier 2-Closed mechanistic investigative-For recurring business questions Tier 3-Proactive-For monitoring business performance against business objectives

II.

Overview

Founded in 1969, The Systems and Services Company (SYSCO), is the leading food service marketer and distributor in North America. SYSCO also handles distribution of supplies and equipment to restaurant businesses across the US. The company has over 420,000 customers that range from huge restaurant chains to small entrepreneurial units. SYSCO employs over 45,000 people including 8000 marketing associates and 9000 delivery associates. In performing our analysis and formulating our recommendations, we analyzed the business model of SYSCO from two perspectives: 2

The Operational Perspective: Consisting of autonomous operating companies and a corporate HQ at Houston. The Conceptual Perspective: Consisting of Marketing associates to push sales and Delivery associates to transport orders with an overlapping relationship management function to facilitate the linkage between supplier and end user.

III.

Problems and Alternatives

Problem 1: Difficulty in determining quantity of software to purchase. SYSCO has 3 approaches under consideration-the bare bones approach, the middle of the road approach and the volume discount approach. The cost spread between the approaches is approximately 1 Million dollars. Alternatives for Problem 1 1a. Adopt the bare bones approach. Positives: This will minimize upfront investment and will enable SYSCO to learn before committing to additional software. The centralization of data analytic functions is a relatively unknown territory for SYSCO. It would be ideal, in such a scenario to restrict access to the data, which is facilitated by the bare bones approach. It also minimizes training expenses. Negatives: The bare bones approach will not allow licensing of the customer intelligence analytical module. Recreating the calculations built into the model is a time consuming process and may ultimately result in an increase in developmental costs. 1b. Adopt the middle of the road approach. Positives: This will enable licensing of the customer intelligence analytical model and more licenses than the bare bones approach. Negatives: The cost difference between this approach and the volume discount approach is not significant. The volume discount approach, on the other hand, gives an additional supply chain module. The customer analytical module on its own will not give an accurate predictive analysis. Accurate analysis will need cross value chain information. 1c. Adopt the Volume discount approach. Positives: This approach is the best from an overall cost perspective as well as an adoption perspective. Business Objects is bound to give a discount on licenses purchased if SYSCO buys everything up front.A larger group of people will be able to view business intelligence data .Moreover,this approach gives SYSCO the opportunity to use the supply chain module along with the customer intelligence analytical module. Negatives: Totally embedding Business Objects BI architecture and implementation will be detrimental in a fluid and developing knowledge area. Business Objects proprietary semantic layer, while groundbreaking, may not support the unique business model of SYSCO.

Problem 2: Lack of clarity in business intelligence delivery The value of business intelligence increases as the delivery of information is embedded in the processes and systems of the enterprise. It is evident that the road ahead has not been charted by SYSCO as far as implementation is concerned. The delivery of the BI platforms to different levels of consumers is presently based on using the connectors and the semantic layer and merely embedding the modules procured. This approach, facilitated using Business Objects BI solution accelerator does not cater for integrated analytics across the spectrum of customer intelligence, relationship issues and supply chain intelligence-major factors for SYSCO.

Alternatives for Problem 2 2a. Have a SYSCO business model centric 3 tired approach to Business intelligence delivery to answer the following: non-recurring business questions, recurring business questions and proactive intelligence delivery. Utilize Business Objects modular platform to obtain Intelligence on demand and develop the unstructured analytics in-house. Positives: This will be appropriate in a business model as used by SYSCO. The customer relations and the supply chain processes are intricately linked and overlapped by the relationship management routine that SYSCO follows. This calls for a cross value chain information feed that can facilitate integrated analytics. A 3 tiered approach to delivery can solve the issue of non recurring questions that arise in such an open adaptive environment as well as deliver structured information on recurring business questions. A third level will intelligently push information directly to end consumers by continuously monitoring on going business performance against business objectives. Negatives: This will require redesigning the delivery platform with the core components of the Business Objects architecture. This is bound to be time consuming and more expensive than merely embedding the modules using the BI solution accelerator. 2b. Focus on working with the Business Objects consultants using their BI solution accelerator to use connectors and semantic layer to embed existing modules. Positives: Cost effective. The BI solution Accelerator will go through the scoping, design, building, deployment and evolution of the purchased modules. The consulting costs will cover this expense. Negatives: Will not facilitate the integrated analytics layer and cross value chain information analysis enabled by the 3 tiered approach. Problem 3: Lack of evaluation of existing BI applications in operating companies. Most of the operating companies already have their own BI implementations. The autonomous nature in which these companies have been operating means that they have an in-depth knowledge of their playing field. The solutions they have individually implemented to obtain competitive intelligence are of immense value and there is a need to at least study them, if not incorporate them in the final implementation process. A current state analysis, desired future state and a gap analysis, at this juncture, will be invaluable. 4

Alternatives for Problem 3 3a. Study the individual implementations of BI and incorporate relevant aspects in the final implementation of the platform. Positives: The individual companies know their turf very well. The applications they have implemented may have a number of elements that are crucial to their operating environment. These elements can make a difference in leveraging the effectiveness of the platform. There is a possibility that costs may reduce. Negatives: The implementation time schedule could be affected. There is an equal chance of overshooting the costs as implementation and incorporation costs can increase. 3b. Implement the Business Objects solution as it is without incorporating elements of the individual applications. Positives: Going ahead with a known solution will reduce implementation risks and cost. It will also ensure uniformity of application across the various operating companies. Negatives: The opportunities that may be available in the local environment will be lost.

Problem 4: Change management issues linked with centralized data mining in a decentralized business model. The business model that SYSCO follows is of a predominantly decentralized nature. The process of Business intelligence implementation calls for centralized architecture and data/information sharing. The success of SYSCO has been predominantly due to the autonomous functionality given to the operating companies. In an effort to leverage and adopt a data-centric approach, the competitive advantages of the decentralized approach should not be compromised. Alternatives for Problem 4 4a. Deploy the three level architecture, which gives structured analytics to the operating companies and integrated analytics to the corporate HQ. Positives: This will sustain the autonomy of the operating companies and at the same time enable the company to take advantage of different levels of analysis Negatives: Implementing such a platform may increase costs and the time for delivery. 4b. Change the power structure to make the business model more HQ centric. Positives: Could take advantage of opportunities that can crop up in different sectors that may be unknown to the individual companies. Negatives: The marketing-relationship management-supply chain business model that SYSCO follows, calls for an approach which has to give due importance and relative autonomy to the operating companies. Disturbing this structure may be detrimental to SYSCO.

4c. Maintain the existing procedure and merely centralize the data repository functions. Positives: The positives of the business model will be sustained and the data can be used for in depth analysis. Negatives: The change management problems may not be solved as access to information is a critical aspect to the power structure. IV. Scenarios

1. Business Objects can be affected by the increasing competition from other BI service providers and go out of business. There is a higher than average likelihood of this happening. BI is a relatively new field and implementation techniques and protocols are still being developed. Business Objects has a modular approach to implementation which may not suit the open adaptive nature of analysis in an ever changing environment. 2. A shift of business functionality from decentralized mode to centralized mode

There is a higher than average likelihood of this happening. The last 10 years have seen a shift to a data centric management approach with the ERP implementation and the datawarehousing initiative. With the BI implementation, there is a high probability of the power structure shifting. 3. Unstable Political and Economic conditions that affect overall supply and demand considerations. There is a high likelihood of this happening. SYSCO in fact had to tide over difficulties during the 2003 when external environmental factors affected supply and demand. (IRAQ War). The need to involve external environmental factors in the predictive analysis process is important. 4. Discontentment among operating companies at information sharing

There is a high likelihood of this happening. Power most often is related with information. The easy access of data and the centralized approach to analysis may cause discontentment among the managers at the operating companies.

V.

Recommendations and Conclusions.(Diagram Reference: APPENDIX 1)

Our recommendations are based on the following considerations: The existing business model of SYSCO has a decentralized structure. The autonomy that this structure provides supports the business processes that are executed by the Marketing associates and the Delivery associates overlapped by an umbrella of relationship management initiatives. 6

The customer analytics and the supply chain are, therefore intricately linked. Deriving intelligence from one without considering the other may not provide accurate results especially in predictive analysis. The business model also derives many inputs from the external environment. The consideration of these inputs is important in deriving answers to non-recurring business questions. The above considerations clearly indicate that the approach considered must have a supply chain module integrated with the customer analytic module. Therefore choosing the volume discount approach is important not only from the perspective of cost discounts and maximum licenses, but also from the point of view of deriving maximum intelligence and performing predictive analytics in this particular environment. With the approach decided, we shift focus to the issue of delivery of the platform. The BI solution accelerator methodology for delivery/implementation of the platform is very modular and scopes, designs, builds and deploys the system to connect the databases and enterprise applications with the dashboards and modules. In the case of SYSCO, this approach does not address unstructured analytic considerations and cross spectrum value chain data mining that is required in such an open adaptive environment. There is a need to deploy a three level delivery approach, keeping in mind the importance of the existing power structure and the importance of capitalizing on the existing BI applications at the operating company level. During the implementation phase, it is critical to integrate the existing BI applications with the Business Objects solution. The three tiered delivery platform will be as follows: Tier 1-Open adaptive investigative level-For non-recurring business questions. Tier 2-Closed mechanistic investigative level-For recurring business questions. Tier 3-Proactive level-For monitoring business performance against business functions. This delivery adds an additional level that the on demand architecture of Business Objects does not provide. This additional level caters to the non-recurring business questions and takes into account inputs from the external environment, crucial in making predictive decisions. This level caters to extended enterprise analysis that takes into consideration cross value chain information that fits with the marketing/relationship management/supply chain model of SYSCO. This analytics layer would be missing if one goes ahead with a mere solution accelerator methodology of implementation followed by Business Objects. The three level architecture also plays the following roles in the case of SYSCO: Helps maintain a semblance of autonomy using the tier 2 and tier 3 solely for the operating companies and the tier 1 at the HQ level. It would facilitate maintenance of competitive advantages of the decentralized business model and smoothen change management in a shifting power structure.

APPENDIX 1: TEAM 5S PROPOSED BI IMPLEMENTATION FOR SYSCO SYSCOs environment


EXTERNAL/INTERNALCUSTOMER/RELATIONSHIPMANAGEMENT/SUPPLY CHAIN

PROPOSED BI DELIVERY PLATFORM FOR SYSCOs BUSINESS MODEL CONSIDERING EXTERNAL INTERNAL ENVIRONMENT

T I E R 1

T I E R 2

T I E R 3

TIER 3 Proactive layerpushes information while doing continuous monitoring

PROPOSED DATA ARCHITECTURE FOR SYSCOs BUSINESS MODEL-CONSIDERING INTERNAL/EXTERNAL ENVIRONMENT

BI IMPLEMENTATION: ON DEMAND-TIER 2 INTEGRATED ANALYTICS-TIER 1

Credits for the diagrams depicted above that have been used to build our BI implementation plan for SYSCO goes to: Software Group, IBM ( WWW.IBM.COM)

Appendix 2: Praxis A shift of business functionality from decentralized mode to centralized mode Higher than average Bad. Does not facilitate centralization of data to the fullest. This would mean increased development and deployment costs in the future. Good. The focus shifts more toward a data centric approach Business Objects getting affected due to increasing competition from other BI service providers Higher than average Neutral-to-Bad. This approach would not require much of vendor support and hence can survive in case of failure of Business Objects.

PRAXIS

Discontentment among operating companies at information sharing

Unstable Political and Economic conditions

Likelihood 1a. Adopt the bare bones approach.

High Neutral-to-Bad. Depends on the amount of data centralized. Higher the data centralization, higher the probability of discontentment among the operating companies. Bad. Incorporates centralization of data across all operating companies.

High Bad. Does not facilitate predictive analysis which might be useful in deriving answers to non-recurring business questions Good. The predictive analysis made by the customer intelligence analytical module might not be accurate due to the absence of cross value chain information. Best. Has both customer intelligence analytical module and supply chain 9

1b. Adopt the middle of the road approach.

Bad. This approach requires more vendor support compared to the bare bones approach and hence BI at SYSCO might be affected if Business Objects gets affected. Bad. Has heavier dependence on Business Objects for technical

1c. Adopt the Volume discount approach.

Bad. Incorporates centralization of data across all operating companies.

Best. Best suited for a centralized environment and this comes at a discounted

module and hence the analysis will be accurate.

price.

Best. Addresses the issues of unstable political and economic conditions better than other models but the development time and cost is the liability here. Bad. This does Bad. Cannot 2b. Work with not address the effectively the business issue of answer the objects displeasure non-recurring consultants among the business using their BI operating questions due solution companies due to the absence accelerator to data sharing. of integrated analytics layer. Good. Good. 3a. Evaluate Individual BI and incorporate Incorporating applications crucial features the BI applications have from the from the invaluable individual BI operating information applications companies which can be into the final implementation. might encourage put to better the managers to use. share data and information. Bad. This does Good. But, to 3b. Implement not address the address these the Business issue of issues at least Objects customer solution as such displeasure intelligence without features among the analytical from individual operating companies due module is applications. to data sharing needed. 2a. Have a SYSCO business model centric 3 tired approach to BI delivery 10

Best. This ensures the autonomy of the operating companies and at the same time facilitates a data centric approach.

Best. Ideal in a centralized environment with a drawback of increased development time and cost. Helps maintain competitive advantages of decentralization

support and hence the software might become obsolete without vendor support. Neutral to Bad. Is dependent on Business object for vendor support and technical support and hence failure at Business Objects affects BI at SYSCO.

Bad. Cannot sustain in a centralized environment without improvement of the software.

Bad. Requires vendor support and a failure of Business Objects might jeopardize the BI at SYSCO.

Neutral-to-Bad. Depends on how well the features from local environment are in sync with each other and also with the central model. Good, provided the software has at least customer intelligence analytical module.

Neutral. Has no effect as the model is developed inhouse and hence no vendor support from Business objects is required. Bad. Requires vendor support and hence BI at SYSCO might be affected if Business Objects gets affected.

4a. Deploy the three level architecture

Best. This ensures the autonomy of the operating companies and at the same time facilitates a data centric approach.

Best. Addresses the issues of unstable political and economic conditions but at the cost of increased development time and cost. Neutral-toBad. Distribution of the power structure might degrade the ability of the model to effectively address these instability issues.

Best. Best suited in a centralized environment with a drawback of increased development time and cost.

Best. Provides the necessary freedom to the operating companies and also takes advantage of the opportunities that can crop up in different sectors unknown to the operating companies. 4c. Maintain the Bad. This does not address the existing issue of procedure and discontentment centralize merely the data among the operating repository companies due functions to data sharing. 4b. Change the power structure to make the business model more HQ centric

Bad. This approach involves distribution of power structure which might not sustain in a centralized environment.

Bad. Is heavily dependent on Business object for vendor support and technical support and hence failure at Business Objects affects BI at SYSCO badly. Neutral. Has no effect as the model is developed inhouse and hence no vendor support from Business objects is required.

Bad. Cannot effectively answer the non-recurring business questions due to the absence of customer intelligence analytical module and supply chain module.

Bad. Cannot sustain in a centralized environment without improvement of the software.

Bad. Requires vendor support and hence BI at SYSCO might be affected if Business Objects gets affected.

The Business intelligence implementation for SYSCO proposed by our team addresses 3 scenarios very effectively. In the probability of the collapse of Business Objects, however, the evolution of BI can be affected quite badly. But this can be countered by training the IT department of SYSCO to provide technical support in case of any application failure and on the systematic usage of the application. An evolutionary plan can be envisaged keeping in mind the basic business model and the information needs of the company. The integrated analytic layer, independent of the Business Objects solution can be utilized as a base to build future technologies on.

11

Вам также может понравиться