The Cole Papers

The key to buying a new system is the specification, not the RFP

After 23 years in the publishing business, the Information Technology Director wore the scars and the stars of more than a dozen major computer system installations. Editorial, advertising, pagination, circulation, accounts receivable, color systems, output management subsystems. You name it; he'd bought it. He'd installed it. He'd replaced it.

He brought this terabyte of experience with him as he walked into the executive conference room to join the kick-off meeting for the new classified and display advertising system. The supplier's team sat to his left, all suited up with laptops poised. The customer team -- his team -- sat together on his right. Without a word, the IT Director faced the white board at the head of the table, picked up a red dry-erase marker and wrote in large, neat upper-case letters: "Windscale." Turning to his audience, he pointed to the board and punctuated each letter with the red marker as he spoke. "What Is Not Down in Specification Costs A Lot Extra," he said to the group. Then he turned and left the room in strategic, deferential silence.

Through this simple, melodramatic gesture, the IT Director communicated to the entire team his belief in the importance of the specification process when launching a new project. In addition, as the project's sponsor, he acknowledged the specification as the unifying document that describes what the system will and will not do -- and ultimately, what his company will and will not pay for.

There's a problem, however, right out of the launch pad. Even if every person sitting at that table agrees the spec is important, there are probably two widely different views as to what a specification actually is and does. Most members of the customer team would say that the specification defines what they want the system to do. The supplier team, on the other hand, is likely to say that the specification describes how the system will work. The distinction between what and how often leads to misunderstandings that can jeopardize the success of the project.

In the customer's eyes, the focus is on needs and requirements. The system search and selection process often begins with a Request For Proposal (RFP) detailing all the things the customer wants in a new system. In the supplier's eyes, the focus is on implementation. The supplier may respond to the customer's RFP with a proposal describing all the features and functionality that their system has to offer.

The customer is talking expectations, while the supplier is talking execution. Disconnects occur whenever the customer feels their RFP very clearly states what they want the system to do, and the supplier feels their system proposal defines exactly what the customer will be getting. As soon as a disagreement arises, each party has a natural tendency to perceive the problem from their own point of view.

The key, therefore, is to recognize from the outset that the specification must encompass the requirements as well as the solutions. The what is certainly important. So is the how. But, to ensure that the specification meets the objectives of both the customer and the supplier, the focus also needs to be on the why. If every individual gathered around that conference table can agree why a particular item appears in the specification, then both parties will begin to see the issues from the other's perspective.

RFP problems
Take, for example, a customer requirement that the system contain a full-featured text processing component. By stating this requirement in the RFP, the customer team believes it is telling the supplier what it wants. In addressing this requirement, the supplier's proposal may state that its system uses Word from Microsoft Corp. of Redmond, Wash., as the text processor. Microsoft Word, with all its built-in capabilities, is how the supplier's system will satisfy the customer's requirement.

The potential problem crops up when trying to interpret why the customer wants a system with a full-featured text processor. From the customer's point of view, this requirement will allow the users to have a fast and flexible means of creating complex formats with the ability to switch to different typographic styles by using a single keystroke.

In the supplier's view, Microsoft Word is an industry-standard document generation tool. When you buy this supplier's system, in other words, you get all the tools that MS Word provides to handle all of your text processing needs. Anyone can learn to write Word macros that can handle all sorts of complex formatting requirements.

But who's going to take responsibility for writing these formats? Who's going to debug them, test them, document them and integrate them? And, when you think back to the IT Director and his Windscale admonition written on that white board, who's going to pay for the time and effort necessary to implement these formats?

As part of the cross-functional spirit that is vital to the success of any project, the supplier and customer team members always should ask themselves why a particular requirement is needed before putting it down in the specification. The focus should be on desired outcomes. This approach will help foster clarity, consensus and a clear set of expectations and deliverables. Using the example above, the requirement for a full-featured text processor could be written in the specification as follows:

"The system will include a text processing program that allows user-definable formatting and programmable keyboard commands. The standard system will include a library of documented, pre-defined formats that can be customized and expanded by users with appropriate authority levels."

In addition to the why rule, here are five corollaries to the Windscale Principle that can be used by the combined supplier and customer teams to help produce good, clear specifications and to avoid communications troubles -- and potential accounts receivable battles -- once the project gets under way.

  • A bad beginning makes a bad ending. This adage, attributable to Euripides in 406 B.C., is certainly relevant today, just as our IT Director understood when he set the agenda for the project team during the kickoff meeting. The specification process is the foundation of the entire project. All subsequent tasks -- including design, development, manufacture, delivery, plant test, integration, training, parallel and live operation -- will suffer if the time is not invested early in the project to write specifications that everyone agrees to.

  • Memories fade a lot quicker than ink. Don't rely on collective memories when defining requirements and crafting solutions. Complex projects often involve many players, some of whom may be long gone by the time the implementation phase rolls around. Verbal promises, assurances and commitments are often difficult to verify, and are wide-open to misinterpretation. The final specification should be viewed by supplier and customer alike as the official rulebook for the project. When disputes arise, as they inevitably will, the team should be able to turn to the specification as the definitive document that captures each deliverable.

  • Keep it simple, succinct, unambiguous, precise. The awkward acronym notwithstanding, this rule will help to ensure that everyone who reads it will interpret the specification the same way. Each item in the specification should express a single thought. Simple declarative sentences should be used. With large projects, several different people often share in the spec'ing duties, so it is important to maintain a consistent writing style.

    Use words such as "will" to state facts, "shall" to state requirements and "should" to state goals. Avoid words that can cause misunderstanding, and that might lead to money disputes and strained relationships down the road. Words like "etc.," "TBD," "and/or," "fast" and "sufficient" do not belong in a specification because they are vague, confusing and not verifiable.

    To be verifiable, the terms in the specification must be specific and quantitative. A sentence such as, "The system will provide rapid, user-friendly access to menus ...," may mean one thing to a customer and something entirely different to a design engineer. Try to choose words that are measurable, objective and straightforward.

  • Know the difference between necessities and niceties. When putting together a specification, you probably do not want to end up with a monster design that will take years to develop, years to install and will probably not be supportable once you've taken it live. A system laden with bells and whistles may be cool, but every one of these add-ons needs to be developed, tested, documented and integrated. This can prove time-consuming and expensive.

    If there is any doubt about whether an item in the specification is really needed, ask yourself: What is the worst thing that could happen if this didn't get included in the final system? If the project team cannot come up with an answer of substance, then the item is probably not really needed. Also, try to weigh the importance of each requirement against the objective of meeting your delivery deadlines and cost constraints.

  • No bull, no bullets. Good specifications should be written in sentence form. Don't over-specify. People often write down everything they can think of, with the hope that the reader will somehow grasp the real requirement. Avoid long, run-on sentences that convey more than a single thought. And, try not to over-use bullets when listing requirements or defining solutions.

    Bulleted lists may be helpful to identify key elements or components of a spec entry, but they should not take the place of simple, positive statements. Bullets by themselves are often vague, confusing and prone to lots of misunderstanding. Therefore, when a bulleted list seems appropriate, be sure to precede each list with an introductory sentence defining the context of the elements that follow.

    Team-based approach
    Remembering our "Costs A Lot Extra" credo, it is interesting to note that the Department of Defense has stated that over-specification is the primary cause of budget overruns on military programs. This highlights the real challenge for project teams chartered to deliver accurate specifications. How do you make sure the specification is complete enough so that everyone agrees to what is being purchased and paid for, yet not so stringently written that the project becomes an expensive, resource-consuming albatross?

    Certainly, it helps to take a team-based approach to the specification process. As with other problem-solving efforts, the first step is to figure out exactly what the team is trying to accomplish, and to identify the priorities and concerns of everyone on the team. Many team initiatives fail when the group jumps right to the solution part of the process. Assuming that everyone on the team has the same orientation and objectives for the final specification can greatly inhibit the group's productivity.

    By taking the time to identify these concerns -- up front -- the chances of misunderstanding will be decreased and the awareness of the multiple dimensions of the specification process will be increased. And, just as important, this dialog will enhance interpersonal relations by sending a message to each team member that his or her opinion is valued. This, in turn, will foster a working environment that is conducive to finding a set of common goals and speaking with a unified voice.

    Finally, as our friend the IT Director illustrates, it is essential for the team's sponsor to buy into the importance of the specification process. It may take time for the team to produce a spec that clearly and accurately captures the scope of the project, but the sponsor needs to realize that this time is well spent. Poorly written or incomplete specs can add many hours of rework time to the project, so the time investment at the beginning of the project is always worthwhile. "Do it once, do it right." "Go slow to go fast." It's been said in many ways. A good sponsor recognizes this and can help facilitate the process by developing a sense of shared leadership and common purpose. During the specification process, and throughout the entire project, the sponsor should take responsibility for guiding the project's mission, building team cohesion, maintaining focus among team members and project stakeholders and resolving conflicts when appropriate.

    The challenge is great, but the rewards are significant. Consistent expectations. Mutual benefits. On-time/on-budget delivery. Happy end-users. Customer fulfillment. Invoices paid in full. Long-term supplier-client relationships. By adhering to the Windscale Principle, and recognizing and respecting the significance of the specification process, the likelihood of project success will be assured.

    -- Peter G. Marsh

    From THE COLE PAPERS, November 1999, Copyright © 1999, All Rights Reserved.

  • Top | ColeGroup.com | Consulting | Cole Papers | NewsInc. | Cole's Store | Miscellanea | Search
    Copyright © 1990-2012, The Cole Group. All Rights Reserved. Contact us.
    Modified date: 07/22/2002, 11:43:34 AM.
    URL: http://www.colepapers.net/tcp.archive/Cole_Papers_99/TCP_99_11/spec.html