I can see why teams like having one number to focus on. It makes progress easier to discuss and gives people a shared goal. The harder question is whether moving that number actually makes the product better.
For Airbnb, nights booked was a sensible choice. It meant guests were finding accommodation and hosts were getting business, without the price and exchange-rate fluctuations that affected revenue. But a team could buy more bookings with huge discounts. Nights would go up, and everyone might celebrate, even if those bookings were losing money.
Facebook’s “seven friends in ten days” gets at something familiar: there’s little reason to open a social app if nobody you know is there. Early friend connections were associated with better retention, and the target gave the growth team something concrete to work toward. The exact seven mattered less than helping people find their circle quickly. Push someone to add seven strangers, though, and you’ve met the target while missing the point.
Slack’s number was 2,000 messages across a team, with no fixed deadline. In a 2015 interview, Stewart Butterfield said 93% of teams reaching that point were still using Slack. That suggested they had given it a real try together. But imagine rewarding teams for sending more messages. Suddenly, a clear update becomes ten separate posts, and everyone has more to read.
These hypothetical shortcuts show the problem behind Goodhart’s Law. Once people are rewarded for the number itself, it can stop telling us what we wanted to know. I’d still track these metrics, but I’d also ask users whether the experience was getting better.
