The Perfect Fit: A Guide to Evaluating and

Purchasing Major Software Systems

By Peter Campbell, September, 2008
A major software package shouldn't be chosen lightly. In this detailed guide, Peter Campbell
walks through how to find software options, evaluate them, make a good decision, and then
purchase the system in a way that protects you.
A smart shopper evaluates the item they want to purchase before putting money down. You
wouldnt shop for shoes without checking the size and taking a stroll up and down the aisle in
order to make sure they fit, would you? So whats the equivalent process of trying on a
software package will size? How can you make sure your substantial software purchase
wont leave you sore and blistered after the cash has been exchanged?

Thats the goal of this articleto provide some guidance for properly evaluating major
software investments. Well walk through how to find potential software options, gather the
detailed information you need to evaluate them, make a solid decision and purchase a
package in a way that protects you if it doesnt do what you hoped it would for you.

Is it A Major Software System?
The evaluation process described here is detailed, so its probably not cost effective to apply
it to every software tool and utility you purchase. How do you know if the package youre
considering is major enough to qualify? Major systems have a dramatic impact on your
ability to operate and achieve your missionthey arent measured by budget, theyre
measured by impact.
To help identify a major purchase, ask yourself:
Will the application be used by a significant percentage of your staff?
Will multiple departments or organizational units be using it?
Will this software integrate with other data systems?
If this software becomes unstable or unusable once deployed, will it have significant
impact on your nonprofits ability to operate?

Giving significant attention to these types of major purchases is likely to save your
organization time in the long run.

Taking Preliminary Measurements
Prior to even looking at available software options, make sure you thoroughly define your
needs and what the application you select should be able to do for you. Nonprofits are
process-driven. They receive, acknowledge, deposit and track donations; they identify, serve
and record transactions with clients; and they recruit, hire and manage employees.
Technology facilitates the way your organization manages these processes. A successful
software installation will make this work easier, more streamlined and more effective. But a
new system that doesnt take your processes and needs into account will only make running
your organization more difficult.
So its critical that, before you begin looking for that donor database or client-tracking
system, you clearly understand the processes that need to be supported and the software
features critical to support that work.
This is an important and complex area that could easily be an articleor a bookin its own
right. We could also write numerous articles that delve into project management, getting
company buy-in and change managementall critical factors in organizational readiness.
However, for the purposes of this article, were focusing on the process of evaluating and
purchasing software once youve already identified your needs and prepped the organization
for the project.

Finding the Available Options
Once you know what you need and why you need it, the next step is to identify the pool of
applications that might fit. An expert consultant can be a huge help. A consultant who knows
the market and is familiar with how the systems are working for other nonprofits can save
you research time, and can direct you to systems more likely to meet your true needs. While a
consultant can be more expensive than going it alone, money spent up front on the selection
and planning phases is almost always recouped through lower costs and greater efficiency
down the road.
If a consultant isnt warranted, take advantage of the resources available to the nonprofit
community, such as Idealware, Social Source Commons, Techsoups forums or NTENs
surveys. Ask your peers what theyre using, how they like it and why. Ideally you want to
identify no less than three, and probably no more than eight, suitable products to evaluate.

Considering an RFP
With your list of possible software candidates in hand, the next step is to find out more about
how those packages meet your needs. This is traditionally done through a Request for
Proposal (RFP), a document that describes your environment and asks for the information
you need to know about the products youre evaluating.
Well-written RFPs can be extremely valuable for understanding the objective aspects of large
software purchases. For example, if you are looking for a Web site content management
system (CMS), questions such as does the blogging feature support trackbacks? or Can the
CMS display individualized content based on cookie or user authentication? are good ones
for an RFP.
What you want from the RFP is information you can track with checkboxes. For example, It
can/cant do this, It can/cant export to these formats: XML, SQL, CSV, PDF, or They
can program in PHP and Ruby, but not Java or Cold Fusion. Questions that encourage
vendors to answer unambiguously, with answers that can be compared in a simple matrix,
will be useful for assessing and documenting the system capabilities.
An RFP cant address all the concerns you're likely to have. Subjective questions like How
user-friendly is your system? or Please describe your support are unlikely to be answered
meaningfully through an RFP process.
Certainly, you can arrange for demonstrations, and use that opportunity to ask your questions
without going through an RFP process. But while the formality of an RFP might seem
unnecessary, there are some key reasons for getting your critical questions answered in
writing:
You can objectively assess the responses and only pursue the applications that arent
clearly ruled out, saving some time later in the process.
A more casual phone or demo approach might result in different questions asked and
answered by different vendors. An RFP process puts all of the applications and
vendors on a level field for assessing.
The RFP responses of the vendor you select are routinely attached to the signed
contract. An all-too-common scenario is that the vendor answers all of your questions
with yes, yes, yes, but the answers change once you start to implement the software.
If you dont have the assurances that the software will do what you require in writing,
you wont have solid legal footing to void a contract.

Structuring Your RFP
RFPs work well as a four section document. Below, we walk through each of those sections.
Introduction
The introduction provides a summary of your organization, mission and the purpose of the
RFP
Background
The background section provides context the vendor will need to understand your situation.
Consider including a description of your organizationfor instance, number of locations,
number of staff and organizational structure, the processes the system should support, and
such technology infrastructure as network operating system(s) and other core software
packages. Include any upcoming projects that might be relevant.
Questionnaire
The questionnaire is the critical piece of the documentyou want to be sure you ask all of
the questions that you need answered. In preparing these questions, its best to envision what
the vendor responses might look like. What will have to be in those responses for you to
properly assess them? Consider asking about:
Functionality. In order to get answers youll be able to compare, ask your questions
at a granular level. Does a CRM support householding? Does a donor database have a
method for storing soft credits? Can multiple users maintain and view records of
donor interactions? Can alerts or notifications be programmed in response to
particular events? Use the results of your business requirements work to focus in on
the functions that are critical to you and your more unusual needs.
Technology specifics. Make sure the software will integrate properly with other
applications, that the reporting is robust and customizable by end users, and that the
platform is well-supported. Ask which formats data can be exported to and imported
from, how many tables can be queried simultaneously and what type of support is
availableboth from the vendor and third parties. Ask for a data dictionary, which a
technical staffer or consultant can review, because a poorly designed database will
complicate reporting and integration. And ask for a product roadmap. If the next
version is going to be a complete rewrite of the application, you might want to rule
out the current version for consideration.
Company information. Think through what youll want to know about the company
itself. How big is it? Do they have an office near you? How long have they been in
business? Are they public or private? Can they provide some documentation of
financial viability? Who are the staff members that would be assigned to your project?
References from similar clients with similar-scope projects can also be very useful.
For more information on this area, see Idealwares article Vendors as Allies: How to
Evaluate Viability, Service, and Commitment.
Pricing and availability. What are their hourly rates, broken down by role, if
applicable? What are their payment terms? What is their total estimate for the project
as described? How do they handle changes in project scope that might arise during
implementation? What are their incidental rates and policies (travel, meals)? Do they
discount their services or software costs for 501(c)(3)s? How long do they estimate
this project will take? When are they available to start?

While its important to be thorough, dont ask a lot of questions you dont plan to actually use
to evaluate the systems. Asking questions just in case increases the amount of information
youll need to sift through later, and increases the possibility that vendors might decide your
RFP isnt worth the time to respond to.
Instructions
Close with a deadline and details about how to submit replies. For a sizeable RFP, allow a
minimum of four to six weeks for a response. Remember that this isnt a confrontational
processa good vendor will appreciate and want to work with a client that has thought
things out this well, and the questionnaire is also an opportunity for them to understand the
project up front and determine their suitability for it. Respect their schedules and give them
ample time to provide a detailed response.
Include an indication as to how additional questions will be handled. In general, if one vendor
asks for clarification or details, your answers should be shared with all of the RFP
participants. You want to keep things on a level playing field, and not give one vendor an
advantage over the rest. You might do this via a group Q&A, with all the vendors invited to
participate in a meeting or conference call after the RFP has been sent to them but well before
they are due to respond. With all vendors asking their questions in the same room, you keep
them all equally informed. Alternatively, you can specify a deadline by which written
questions must be submitted. All participants would then receive the questions and answers.

Evaluating the Answers
Once you receive RFP responses, youll need to winnow down your list to determine which
packages youd like to demo.
If you asked straightforward, granular questions, youll now reap the benefit: you can set up a
comparative matrix. Create a table or spreadsheet with columns for each vendor and rows for
each question, summarizing the responses as much as possible in order to have a readable
chart. You might add columns that weight the responses, both on the suitability of the
vendor's response (e.g. 1, unacceptable; 2, fair; 3, excellent) and/or on the importance of the
question (for instance, some features are going to be much more important to you than
others).
Going through the features and technology sections, youll see the strong and weak points of
the applications. In determining which fit your needs, there will likely be some trade-offs
perhaps one application has a stronger model for handling soft credits, but another has more
flexible reporting. Its unlikely that any will jump out as the perfect application, but youll be
able to determine which are generally suitable, and which arent.
For example, if youre looking for software to manage your e-commerce activities, inventory
management might be a critical function for you. If a submitted software package lacks that
feature, then youll need to eliminate it. As long as you understand your own critical needs,
the RFP responses will identify unsuitable candidates.
You might rule out a vendor or two based on what the RFP response tells you about their
availability or company stability. Take care, though, in eliminating vendors based on their
RFP pricing information. RFP responses can be very subjective. Before determining that a
vendor is too pricy based on their project estimate, dig deeperother vendors might be
underestimating the actual cost. If you feel you have a solid grasp on the project timeline, use
the hourly rates as a more significant measurement.
The RFP responses will tell you a lot about the vendors. Youre asking questions that are
important to your ability to operate. Their ability to read, comprehend and reasonably reply to
those questions will offer a strong indication as to how important your business is to them,
and whether theyll consider your needs as the software is implemented and into the future. If
they respond (as many will) to your critical questions with incomplete answers, or with stacks
of pre-printed literaturesaying, in effect, the answers are in here--then theyre telling you
they wont take a lot of time to address your concerns.
Keep in mind, though, that a weak sales representative might not mean a weak vendor,
particularly if they're representing a product that comes recommended or looks particularly
suitable on all other fronts. It's acceptable to reject the response and ask the vendor to
resubmit if you really feel they have done you, and themselves, a disservicebut temper this
with the knowledge that they blew it the first time.

Trying It All On for Size
At this point the process will hopefully have narrowed the field of potential applications
down to three-to-five options. The next step is to schedule software demos. A well-written
RFP will offer important, factual and comprehensive details about the application that might
otherwise be missed, either by too narrow a demo or by one the vendor orchestrates to
highlight product strengths and gloss over weaknesses. But the demos serve many additional
purposes:
Evaluating look and feel. As good as the specs might look, youll know quickly in a
demo if an application is really unusable. For instance, an application might
technically have that great Zip code lookup feature you asked about in the RFP, but it
may be implemented in a way that makes it a pain to use. Prior to the demo, try to
present the vendors with a script of the functions you want to see. It can also be useful
to provide them with sample data, if they are willingevaluating a program with data
similar to your own data will be less distracting. Be careful not to provide them with
actual data that might compromise youror your constituents'privacy and security.
The goal is to provide a level and familiar experience that unifies the demos and puts
you in the driver's seat, not the vendor.
Cross training. The demo is another opportunity for the vendor to educate you
regarding the operating assumptions of the software, and for you to provide them with
more insight into your needs. A generic donor management system is likely to make
very good assumptions about how you track individuals, offer powerful tools for
segmentation and include good canned reports, because the donor-courting processes
are very similar. But in less standardized areasor if you have more unusual needs
the model used by the software application can differ dramatically from your internal
process, making it difficult for your organization to use. Use the demo to learn how
the software will address your own process and less conventional needs.
Internal training. Even more valuable is the opportunity to use the demos to show
internal staff what theyll be able to do with the software. Demos are such a good
opportunity to get staff thinking about the application of technology that you should
pack the room with as many people as you can. Get a good mix of key decision-
makers and application end-usersthe people who design and perform the business
processes the software facilitates. The people who will actually use the software are
the ones who can really tell if the package will work for them.

Making the Decision
With luck, your vendor selection process will now be complete, with one package clearly
identified as the best option. If key constituents are torn between two options or unimpressed
with the lot, senior decision-makers might have to make the call. Be careful, however, not to
alienate a group of people whose commitment and enthusiasm for the project might be
needed.
If none of the applications you evaluated completely meets your needs, but one comes close,
you might consider customizations or software modifications to address the missing areas.
Note that any alterations of the basic software package will likely be costly, will not be
covered in the packaged documentation and help files, and might break if and when you
upgrade the software. Be very sure there isnt an alternate, built-in way to accomplish your
goal. If f the modification is justified, make sure its done in such a way that it wont be too
difficult to support as the software is developed.
Before making a final decision, you should always check vendor references, but take them
with a healthy grain of salt. An organizations satisfaction with software depends not only on
how well it meets their needs, but how familiar they are with their optionsthere are a lot of
people who are happy using difficult, labor-heavy, limited applications simply because they
dont know there are better alternatives.
If you still have a tie after RFPs, demos and reference checks, the best next step is to conduct
on-site visits with an existing customer for each software package. As with demos, bring a
representative group of management, technical staff and users. Assuming the reference can
afford the time to speak with you, the visit will highlight how the software meets their needs,
and will give you a good, real world look at its strengths and weaknesses. Youll also likely
walk away with new ideas as to how you might use it.

Signing on the Dotted Line
Youve selected an application. Congratulations! You might be tired, but you arent finished
yet. You still need to work with the vendor to define the scope of the engagement, and an
agreement that will cover you in case of problems. A good contract clearly articulates and
codifies everything that has been discussed to date into a legally binding agreement. If, down
the road, the vendor isnt living up to their promises, or the software cant do what you were
told it would do, then this is your recourse for getting out of an expensive project.
Contract negotiations can take time. Its far more dangerous to sign a bad contract in the
interest of expediency, though, than it is to delay a project while you ensure that both
partiesyou and the vendorcompletely understand each others requirements. Dont start
planning the project until the papers have been signed.
A software contract should include a number of parts, including the actual agreement, the
license, the scope of work and the RFP.
The Agreement
This is the legal document itself, with all of the mumbo jumbo about force majeure and
indemnity. The key things to look for here are:
Equal terms and penalties. Are terms and penalties equally assessed? Vendors will
write all sorts of terms into contracts that outline what you will do or pay if you dont
live up to your end of the agreement. But theyll often leave out any equivalent
controls on their behavior. You should find every if this happens, customer will do
this clause and make sure the conditions are acceptable, and that there are
complementary terms specified for the vendors actions.
Reasonable cancellation penalties. If there are penalties defined for canceling a
consulting or integration contract, these should not be exorbitant. Its reasonable for
the vendor to impose a limited penalty to cover expenses incurred in anticipation of
scheduled work, such as airfare purchased or materials procured. But unless this is a
fixed cost agreement, which is highly unusual, dont let them impose penalties for
work they dont have to dofor example, for a large percentage of the estimated
project cost.
Agreement under the laws of a sensible state. If the vendor is in California, and
youre in California, then the agreement should be covered by California laws rather
than some random other state. In particular, Virginias laws highly favor software
companies and vendors. In most cases, you want the jurisdiction to be where you live,
or at least where the vendors headquarters actually are.

The Software License
The license specifies the allowed uses of the software youre purchasing. This, too, can
contain some unacceptable conditions.
Use of your data. A software license should not restrict your rights to access or work
with your data in any way you see fit. The license agreement will likely contain
conditions under which the software warranty would be voided. Its perfectly
acceptable for a commercial software vendor to bar re-engineering their product, but
its not acceptable for them to void the warranty if you are only modifying the data
contained within the system. So conditions that bar the exporting, importing,
archiving or mass updating of data should be challenged. If the system is hosted, the
vendor should provide full access to your data, and the license should include
language providing that client shall have reasonable access for using, copying and
backing up all customer information in the database. There should be no language in
the contract implying that the vendor owns your data, or that they can use it for any
additional purposes.
Responsibility avoidance. Software warranties should not include blanket software
provider is not responsible if nothing works statements. This shouldnt need to be
said, but, sadly, there are often warranty sections in license agreements that say just
that.
Back doors. The license should not allow for any post-sale reversals of licensing,
such as language stating that the contract will be void if the customer uses the
software in perfectly reasonable ways they dont anticipate. For instance, if you want
to use the CRM functions of your donor database to track contacts that arent potential
donors, you shouldnt sign a contract limiting use of the software to fundraising
purposes. Also, there should not be any back doors programmed into the
application that the vendor can maintain for purposes of disabling the software.

The Scope of Work
The Scope of Work (SOW) describes exactly what the project will consist of. Its an
agreement between the vendor and the customer as to what will happen, when, and how long
it will take. Good scopes include estimates of hours and costs by task and/or stage of the
project. The scope should be attached as a governing exhibit to the contract. Usually, this is
negotiated prior to receiving the actual contract. By having it attached to the contract, the
vendor is now legally obligated to, basically, do what they said they would do.
The RFP
Like the Scope of Work, the RFP should also be attached as a governing document that
assures that the software does what the vendor claimed it would.
In Conclusion
For big ticket purchases, it's well worth having an attorney review or assist in negotiations.
Keep in mind that the goal is to end up with a contract that equally defends the rights of both
parties. True success, of course, is a solid contract that is never revisited after signing.
Litigation doesn't serve anyone's interest.

Bringing It Home
Theres a lot of talk and plenty of examples of technology jumpstarting an organizations
effectiveness. But if someone were to do the tally, there would probably be more stories of
the reverse. All too often, organizations make decisions about their software based on
uninformed recommendations or quick evaluations of the prospective solutions. Decisions are
often based more on expediency than educated selection.
Rushing a major investment can be a critical error. Learn about the available options,
thoroughly assess their suitability to your needs and prepare your staff to make the most of
them. Then, sign a contract that protects you if, after all else is done, the application and/or
vendor fails to live up to the promises. Finding the right application and setting it up to
support, not inhibit, your workflow is a matter of finding something that really fits. You cant
do that with your eyes closed.

For More Information

Vendors as Allies: How to Evaluate Viability, Service, and Commitment
An Idealware article on how to think specifically about the less-concrete aspects of software
selection.

How To Find Data-Exchange-Friendly Software
An overview of how to ensure you're going to be able to get data in and out of a software
package. (For much more detailed considerations, see our Framework to Evaluate Data
Exchange Features.)

Peter Campbell is the director of Information Technology at Earthjustice, a nonprofit law
firm dedicated to defending the earth, and blogs about NPTech tools and strategies at
Techcafeteria.com. Prior to joining Earthjustice, Peter spent seven years serving as IT
Director at Goodwill Industries of San Francisco, San Mateo, and Marin Counties, and has
been managing technology for non-profits and law firms for over 20 years.
Robert Weiner and Steve Heye also contributed to this article.