How to Use the Requirements Creation Process to Improve Project Estimates

Member Submitted

Estimation can be one of the most difficult parts of a project, especially early in the project. Fortunately, the requirements creation process can lend insight into the length and scope of a project.

Estimation can be one of the most difficult parts of a project. Important questions must be asked in order to form the right figures and plans. How long will the project take? How many resources will it consume? Consultants may also ask the following question: What is the appropriate amount to bid on this project? These questions are not easy to answer at the outset when one generally has only a vague idea of what will be required throughout the project.

The good news is that there is a fairly simple way to improve project estimation and, consequently, the bidding process. Most people do not realize that the requirements creation process can lend insight into the length and scope of a project. Let me give you an example of how this method works and then explain how you can implement it within your own company.

The Story
Back in 1992, I was working for a consulting company named The Kernel Group (TKG). During this time, I was put in charge of porting Tivoli's software from Sun’s Solaris operating system to IBM’s AIX operating system. The project was to be done under a fixed bid, and consequently, we at TKG knew that estimating the effort required to port the code was of paramount importance.

I looked at the code with a coworker of mine, and he came to the conclusion that if Tivoli didn't make the project hard for us in some unspecified way, we could port the million or so lines of code in about a weekend. I told him that he was nuts–that it would take at least a week, maybe even two. We jointly decided that we probably ought to call it three weeks just to be safe. We also decided, rather smugly, not to report our padding of the schedule to evil upper management.

As a result, evil upper management drastically expanded the project bid to $325,000, and my coworker and I thought that this was a ridiculously high price. We believed that we were gouging the customer and that they would never accept it.

Yet they did accept it, and once the project began, we proceeded to discover how truly terrible we as software engineers were at the task of project estimation. To make a long story short, the porting schedule expanded to exceed our original estimate and we consumed not only all of the $325,000, but a whole lot more on top of it.

The Formula
Now our consulting company was religious about tracking employee time on a per-project basis, and so we broke every project into phases: requirements/specification, design, coding, testing, debugging, documentation, training, etc. This project was no different in that respect; we broke it down into its respective phases as well.

Just before we started working on the project in question, I read a book called Practical Software Metrics for Project Management and Process Improvemen by Robert B. Grady. (By the way, this is a truly fabulous book that I would highly recommend to anyone who is managing software development projects.) According to the book, one of Grady’s rules of thumb is that 6-8% of every software project is usually eaten up in the requirements/specification phase.

One of the conclusions that Grady comes to in his work is that you can use this fact to estimate total project size. In other words, if it took 60 hours to do the specification, that's probably 6% of the job and the job will be 1000 hours. Following such logic, a six hour specification implies a 100 hour job. Since the specification always comes first

About the author

AgileConnection is a TechWell community.

Through conferences, training, consulting, and online resources, TechWell helps you develop and deliver great software every day.