I was introducing some ideas I’ve learned recently about throughput management to a friend of mine who is on a large project. The question came up, how to make throughput a useful metric when there are some very small features that go through the system and others that can take months with a large team.
Let’s review a bit: throughput is the measurement of number of items you process through a system over a period of time. So you would say that you did fifteen features and bugs last month, and ten the month before.
If you had no control or process in place to deal with the fact that a new customer might want its killer feature, which is months of work for you, then this metric will quickly become meaningless. What to do?
There is a simple way to handle this and a complicated way. I suppose the complicated way will be what’s needed for my friend’s project, but the simple way is good enough for my project. [Read more…]