|
For every design and development project, Menlo High-Tech Anthropologists® transform their designs into story cards. Each week, the developers estimate the cards we will propose for prioritization by the project stakeholder. We do this to share context, validate the card, identify risk/complexity, and of course, assign a preliminary cost to completing the functionality. On one of our larger projects, the project stakeholder authorized two pairs of developers to spend 24 hours to estimate a large number of cards to facilitate release planning. This was a rare request and a great opportunity to run a few experiments.
How did we schedule the authorized 24 hours?
We broke it down into six sessions, each being four hours with two pairs of developers. We also made sure no developer would be in back to back sessions, avoiding the repetitive nature of this process. We found this to be very effective. One “shift” of developers would estimate cards from 9am to 1pm (we even fed them!). Then, a second “shift” would estimate from 1pm to 5pm. As facilitators, we saw a lot of benefit from shorter shifts. Developers would have a chance to stop estimating for the day, take a break, and go back to doing development work; meanwhile, two new pairs of developers would come in and we’d continue estimating the next cards.
We realized we hadn’t struck the right balance between discussing the card in depth and providing an estimate. In order to estimate all the currently written cards in the 24 hours, we could only allot 8 minutes to every card discussion and estimation. To accomplish this, we put a countdown timer on the screen, keeping us aware of how long we had discussed each card. The goal of this was not to say “after 8 minutes we must be done,” but to keep our conversations focused on getting an estimate in a timely manner. The timer made it clear to us how much time we had spent conversing. It enabled us to ask, “is this conversation leading us to an estimate?”. We’d keep discussing if it was getting us to the estimate. If we needed more information to provide an estimate, we’d stop and document any questions.
We ended up getting through roughly 250 cards in the 24 hours we had scheduled, and learned a lot about how we can best run estimation sessions for this project. A successful experiment!
Have an experiment you'd like featured in the Menlo Bits?
Email us!
|