For Airbnb, nights booked was chosen as a core metric since it serves as a proxy for how both parties (hosts and guests) are using the service and whether successful transactions occur. This is important as Airbnb operates as a two-sided marketplace, meaning its success will depend on attracting hosts willing to list their properties and guests willing to book them. A team could game this by discounting longer stays which increases the number of nights booked without necessarily improving the user experience.
With Facebook’s early growth team, they focused on getting new users to 7 friends in 10 days after discovering users who met this threshold were more likely to stay active. This metric served as a proxy for users establishing meaningful social connections and staying updated through Facebook’s networks. However, a team could game this metric by incessantly recommending friend requests or suggest connections that the user does not know. This would in turn make the social experience less personal.
Slack identified 2000 messages per team as a key threshold after noting that 93% of teams that reached it continued using the product. This metric is a proxy for teams integrating Slack into their main communication workflows. Though, if a team optimizes for message volume, it could create unnecessary conversations and notifications which would cloud communication channels.
With these examples, it’s clear that their North Star metrics are chosen for their ability to reflect meaningful user behavior and long-term engagement. However, Goodhart’s Law shows that when these metrics become targets, teams can end up optimizing for the numbers rather than the value they were originally intended to measure.
