Downloads don’t matter until day 7 does
If people leave in their first week, more installs only make the leak bigger. The one number I check first.
When growth feels slow, the first instinct is to get more installs: more ads, more posts, more launches. Before spending a cent, check one number: how many new users are still active on day 7?
Why day 7
Day 1 tells you whether your onboarding works. Day 30 tells you whether your product works. Day 7 sits in between, and it’s early enough to act on. For a subscription app it also lines up with most trials, so it predicts revenue better than any download chart.
How to read it
Look at your day-7 retention by weekly cohort: all the users who installed in the same week. A single average hides too much. You want to see whether each new week does better than the last, and whether a release or a channel changed it.
- If it’s rising week after week, your product changes are working. Now more installs will pay off.
- If it’s flat, more installs will bring more of the same. Fix the first week first.
- If one channel’s cohorts are much lower, that channel is bringing the wrong people.
Where the first week usually leaks
- Too much before the first win. Every screen before something useful happens costs you people.
- No reason to come back tomorrow. A reminder, a streak or a saved result gives people a reason to open the app again.
- The wrong promise. If the ad or the store listing promised something the app doesn’t do first, people leave fast.
More installs into a leaky first week is the most expensive way to grow.
Try this week: pull day-7 retention for your last eight weekly cohorts. Put them in a row. The trend tells you whether to spend on growth or on the product.