We use ‘perfect pair weeks’ for our mid term estimates. It’s a high enough level that it doesn’t involve getting to deep into the complexity (making estimation processes very long) but to a level where we feel we can have confidence in the estimate being something we are happy for rest of the business to base their schedules and communications on
We don’t actually communicate the perfect pair week figures. We add an amount of time to the estimates as we believe that developers actually spend 60% of their time sat writing code for our stories. We don’t think this is an issue as the rest of the time is spent doing all the other tasks we think are important. For example training and development, standups (~10 minutes each morning) and holidaying
I wanted to question this 60% to ensure it was accurate. I therefore went to our time tracking system (quick and easy) and recorded the time for each developer in my team to see how much had been booked to stories, and how much time had been booked to other things…
I then totalled this and… Hurrah, 57%
I was pleased as it meant I could justify the additional time we put in our estimates and be confident in its accuracy, while reassuring myself it wasn’t just buffer
I hope this helps. Feel free to comment, share and discuss