A couple of weeks ago someone shared a picture illustrating Goodhart's law in our Slack channel.
The law states: "When a measure becomes a target, it ceases to be a good measure." I love that statement but I find the examples in the picture to be a bit misleading. It reads as if goal-setting is a bad thing, whereas Goodhart's law is a simple way to caution people against the side effects of goal-setting.
Let's go with the picture first. If you're using signups as a measure of success, then you might see some funny strategies emerging. Your team could make your product free, or even pay people to create an account. You'll end up with more signups, but not necessarily more customers in the long run—probably not the result you were going for.
But here's what's missing from that picture: the answer to this problem is not to remove targets. But rather to see if you can pair a metric with another one that will act as a guard rail:
- Increase signups by 50%
- Keep week-4 retention of new users above 40%
You keep the signups target but add a retention goal to control the quality of the new leads. Now, your teams will have a simple way to experiment with wild ideas while ensuring that they keep the business healthy.
This principle works really well with OKRs. Pairing your Key Results can help control the side effects of setting your team's focus on a narrow set of targets.
- Here's an excellent post about unintended consequences and Goodhart's law
- More from us on the risks of Key Results obsessions