Goodhart’s Law: When a measure becomes a target, it stops being a good measure.
If you’ve been in engineering leadership for any length of time, you understand that our field has a long history of bad metric targets that have driven engineers to find creative and unproductive ways to game the system.
At the same time, there’s value for us as leaders to have some measurable way to know what’s happening on our team.
But here are several ways measures go wrong, and what you can do instead:
🔵 Tracking measurements for individual team members drives people to avoid imagined personal judgment rather than improving their useful skills. Instead, evaluate measurements only at a team, department, or organization level.
🔵 Highlighting a single metric as a singular point of interest encourages abandonment of other values. Instead, make sure your highlighted measurements come in pairs, as “left and right oars,” so you can see when one value is being sacrificed to achieve another.
🔵 Tracking anything that might imply a measurement of quality can create fear. Doing so without telling the team or without explaining how the measure will be used in a non-judgmental way magnifies that fear. Before starting any measurement like this, first make sure your team trusts your honesty and integrity. Then, be sure to explain how the metric will be used to support and benefit the team members.
🔵 When you have a goal you’re trying to reach, it’s often crucial to have measurements in place to track your progress. In some cultures the measurements can start to become more important than the goal itself. Design your system so that progress toward the goal is easier than the effort and energy required to manipulate the metrics.
There’s a lot more one could say about metrics and measures in an engineering organization, but I hope these high-level ideas will stimulate your thinking. If you’d like to explore your situation further, let’s talk. Visit stevedwire.com/talk to start the conversation.
Leave a Reply