I was at a client's office yesterday for the final review meeting of a two-week Use Case exercise. The particular area of the operations that we were pulling out into Use Cases had been the subject of various patchy projects within the organisation over a couple of years. As far as I can tell some of that work had been really good and laid the groundwork for the Use Cases, but it was not organised into any kind of overall picture or process and there was a certain lack of confidence in the previous projects. The concept of the Use Case approach was really quite radical for them, and the timelines we were talking about to draw them out sounded crazy compared to what had gone before.

This is something I come across often with Use Cases when looking at a business area which has been burnt with a 'traditional' approach before. Often good functional requirements and designs do result from these requirements gathering phases but they are just not organised, coherent and in a form that everyone understands and can sign off.

With Use Cases, everyone is brought along together. The business, IT, stakeholders - they are all involved right from the start and throughout. And this works because Use Case definition phases work best when they are short, sharp and tightly controlled. For any specific 'group' of business operations typically I like to work on a 2-4 week turnaround of a Use Case set. I also like to focus on User Level Use Cases - IT can worry about subfunction ones later. It's more important to focus on what Use Cases are good at. Don't try and do implementation, don't try and do logic, don't try to design business processes... focus on goals and the steps to get there.

And the end result? At the conclusion of the review meeting yesterday one of the key business representatives announced that they had achieved more in the last two weeks than they had in the last two years. Possibly a little unfair to what had gone before, but does show the degree of confidence that can be instilled by following a good requirements approach.