Airbnb’s steering metric focused on nights booked. A booked night means that both a guest found a place they wanted and a host successfully earned money, so this metric captures the value creation on both sides of the marketplace. Under Goodhart’s Law, a team could game the metric by optimizing just for booking volume, such as discounting prices or encouraging lower quality listings just to increase nights. That could grow the metric while making the marketplace worse.
Facebook’s early growth team focused on getting a new user to 7 friends in 10 days, and the reason they did this was because at 7 friends in 10 days, users were likely to experience the product’s social value and stick around. The metric could be gamed by pushing excessive friend recommendations or encouraging low quality connections with people you aren’t really close to. You might hit 7 friends while creating a less meaningful or connected network.
Slack used 2,000 messages per team as their North Star metric. The idea behind this is that once a team exchanged 2,000 messages, it had really tried Slack, and at that point would be likely to keep using it. This metric was a proxy for Slack becoming part of a team’s actual workflow. However if you optimize too much on the volume of messages, that could reward unnecessary messages, notifications, spam, and less meaningful chatter rather than better communication.
The common lesson across these companies is that a good North Star metric should measure experienced value, not activity for activity’s sake. Goodhart’s Law means teams also need guardrail metrics so improving the North Star does not damage the thing it was originally meant to represent.
