The tools that watch how engineering teams work have got a great deal better in the last few years, and they are not going to get worse. How long your changes wait for review. How long you take to review other people's. How many changes you touch, how often you are the one blocking, how much of what goes out under your name you wrote and how much you approved.
Some of that is legitimate. A team genuinely should know where work gets stuck, and pointing at a queue is better than blaming a person.
But plan for a working life where a summary of your activity exists, updates itself, and is read by somebody who does not sit near you and has never watched you work. That is coming to most places, quietly, usually without an announcement.
Two things follow, and I would rather you heard them before it arrives.
The first is to find out what is counted where you work. Ask your manager what report they see about the team and what is in it. You are being described by those numbers whether or not you ever look at them, and it is a strange kind of self neglect not to know the description.
The second is to refuse to let the measure become the work. Every proxy can be fed. You can cut your review time by approving without reading. You can raise your change count by splitting one piece of work into six. The person who plays that game wins two quarters and loses the thing the number was standing in for, and everybody around them can tell, including the people who cannot yet prove it.
And hold on to this. Those systems can count throughput. They cannot see judgement. The most valuable hours of your year are likely to be a conversation that changed a plan before it became code, and an afternoon spent unsticking somebody new. Neither leaves a trace in any report.
So keep a short record of those in your own words, and bring it when the numbers start doing the talking. Measurement is arriving either way. Make sure there is a sentence about you that no graph can produce.
– Asael Shinder
Top comments (0)