A metric that cannot be gamed
Unit 5 of 8 · Lesson 1 of 4Index
You propose a goal for next quarter: 20 percent more weekly active users. Dev does not look up. “I can do that by Thursday. A push notification that says your account has a problem.” He is right, and that is the point.
Every metric has an exploit. Find it first.
How Facebook handled it
Facebook's growth team wanted one number that predicted whether a new user would stay.
It found that people who connected with about seven friends within ten days were far more likely to stick around, and made that the growth team's target. It worked because it tracked the moment people got value, not just activity.
The best metrics measure the moment a user gets value. Sessions can be inflated. Value is much harder to fake.
Doing it with AI, and where it breaks
Ask a model for ten candidate metrics, then for each, how a cynical team would hit the target without helping a single user.
It ranks theoretical abuses above the boring one that actually happens, like a team quietly redefining what counts as active.
Here are ten candidate metrics. For each, describe the cheapest way a team under pressure could hit it without delivering any user value. Then rank them by how likely that is at a 40 person startup.
Build it
One primary metric, one counter metric, and a written exploit analysis for both.
The worked solution
Stuck, or want to compare?
There is a worked solution. Try the build first. Reading it before you have attempted anything is the fastest way to learn nothing.
Checkpoint
If you did the build, these take two minutes. If you cannot answer one of them, that is the part to go back to.
Moving on marks this lesson complete. Finish the build first, it is the part that counts.