There is a particular kind of dashboard that should worry a founder more than a bad one: the dashboard where everything is green. Installs up. Cost per install down. Volume climbing week over week. And underneath it, almost invisibly, a business that is not getting healthier. The metrics moved. The economics did not.
This happens because the easiest numbers to move are rarely the ones that matter, and they are almost always the ones on the report.
Volume is the most seductive number, and the most misread
Install count and cost per install respond instantly to spend, which makes them feel like control. Push budget, watch the line move. But an install is a beginning, not an outcome — and the cheapest install is frequently the least valuable one. Optimising for the number that moves fastest tends to buy exactly the users least likely to stay.
The same trap applies to a rating read as a single figure. A storefront average can hold steady while the recent reviews — the ones a prospective user actually reads — turn sharply negative. The headline number lags the reality it is supposed to describe.
The number that matters sits one step downstream
Cost per install is not the question. Cost per engaged user is closer. Cost per retained, paying user is closer still. The distance between cost-per-install and cost-per-quality-user is where margin is either made or quietly lost — and that distance only becomes visible when acquisition is measured against what happens after the install, not at the moment of it.
This is also where reputation and acquisition turn out to be the same equation seen from two ends. A higher store rating lowers the effective cost of every install without raising spend, because it lifts the conversion of traffic already being paid for. Treating the two as separate line items is how both get optimised against each other.
What to track instead
The honest metrics are slower and less flattering. Cohort-level retention shape, not just whether retention exists. The ratio of lifetime value to acquisition cost, watched over time rather than at a single snapshot. The trajectory of recent review sentiment, which moves before the headline rating does. None of these light up a dashboard the way install volume does. All of them tell you whether the growth is real.
Growth that looks good on a dashboard and growth that compounds in the bank are not the same thing — and the distance between them is usually one honest metric the report left out.
If your acquisition numbers look healthier than your economics, start a private conversation →