Take a run up

The Extreme Programming corollary practice of Negotiated Scope Contract advocates fixing time, cost and quality while scope is negotiated on an ongoing basis. I am currently working on a contract which employs this practice. There is a contract per iteration which fixes time. The composition of the team is fixed per iteration therefore fixing cost. The quality is fixed by adhering to coding standards and ensuring a junit test coverage > 85% is maintained with the outstanding 15% corresponding to functionality that is unlikely to break.

However, before initiating the project the client was insisting on a cost for the project for budgeting purposes. At that stage a user story writing workshop had been completed so we had a large number of user stories but without story point estimates or any idea of the team's velocity. Rather than take a guess, the client was persuaded to fund iterations 0 and 1 so that the team could estimate all the user stories in story points and derive an actual velocity based on the user stories implemented and accepted by the onsite customer. A projected cost for the project was calculated by dividing the total number of story points for the remaining user stories by the team's velocity and then multiplying by the resource cost of the team members.

The client decided that the investment in iterations 0 and 1 was worthwhile because the subsequent estimates to size the overall project would be empirical, i.e. based on a measure of what had already been achieved by the team in iteration 1. The client recognised that having empirical estimates would provide a more realistic cost calculation, which in turn gave the client more confidence when budgeting.


Tags: agile

posted by Simon Baker @ permalink
blinklist : del.icio.us : digg : furl : simpy : spurl : reddit