This was a fun talk from the 2013 ALM Summit, back in January. I didn’t even know it was online until someone tweeted it :).
The ideas expressed here are around the pragmatism of using estimation as a technique, but not getting lost in the weeds. In other words, admit that we need some process and understanding of what we will be doing, but don’t value process over software.
I share some techniques my team and use at Microsoft to keep things understood and on track. Hope you enjoy it.
From the Abstract
Software development is often fundamentally different than the analogies used to describe it. Accordingly, estimating what software will be done and by when can be an exercise in frustration. Yet, some teams seem to just "get it", making accurate predictions of cost or effort. Other teams struggle with excruciatingly detailed planning meetings or with the results of not estimating at all. Explore estimation for software development from several angles including: empirical planning, time-based estimation, relative comparison, and others. Several effective and proven estimation techniques will be explored, expanding your team’s capabilities beyond hoping for the best.