I feel that the best growth metrics try to capture the moment a product becomes genuinely useful. Airbnb wanted to know whether people were actually finding places to stay, Facebook wanted users to build enough of a social network to keep coming back, and Slack wanted teams to make messaging part of their daily workflow. Each company found a measurable behavior that seemed to predict long-term engagement and built its growth strategy around it. I feel like the key here is that they are really good indicators as opposed to goals, because none of these necessarily directly mean a successful product. You could have a social network where everyone is friends with each other but it’s tiny, however, the idea of adding friends indicates the network effect that they were excited about.
The catch related to Goodhart’s Law is that optimizing for these numbers can easily become disconnected from what made them valuable in the first place. Airbnb could drive bookings with unsustainable discounts, Facebook could get people to add friends they don’t care about, and Slack could encourage more messages without improving communication. In each of those cases, the metric could go up without the product necessarily getting better. I think a metric is useful because it reflects something you care about or a core part of how you want your product to work, but once you start chasing the number itself, you risk losing sight of the bigger picture.
