Rough benchmarks, honestly labelled
Averages across consumer categories sit around 25 to 35 percent on day 1, 10 to 15 percent on day 7, and 5 to 8 percent on day 30. Good apps roughly double those. Numbers vary widely by category and country, so treat them as a sanity check, not a target. The useful comparison is your own cohort last month against this month, and since March 2026 App Store Connect gives you peer benchmarks for your category directly, so the agent can tell you where you sit without guessing.
Pull day 1, 7 and 30 retention per weekly cohort from App Store Connect and Play Console for the last 12 weeks, and the category benchmark from App Store Connect. Show me a table of cohorts against the benchmark and tell me in one sentence whether we are above or below median on each day.
Retention is now a ranking signal, which changes the math
Both stores stopped treating download spikes as a quality signal on their own. Apple's ranking weighs retention and engagement after install, and Google's recommendation systems are increasingly driven by post install behaviour. This means a retention fix is also a discovery fix: an app that keeps users climbs for the same keywords without touching the listing. It also means bought installs that leave on day three can hurt rank, which is one more reason the paid loop waits for the retention loop.
The day 3 leak
Most apps lose more users between day 1 and day 3 than in the following four weeks combined. Look at what happens on day 2 in your app. Very often the answer is nothing: no reason to come back, no push permission, no second win. The users who declined the push prompt on day 1 are usually the ones who never return. Finding the leak is a reading job, and it is the first thing the agent does with your analytics.
Find the single largest drop in our retention curve between two consecutive days. For the users who left there, tell me what they did in their last session and what they did not do that retained users did. Name the one event that best predicts day 30 retention and report how many new users reach it.
Fix one: ask for push after the first win
Asked at launch, most users decline the push prompt and the app has no way back to them. Asked after the first win in the app, with a sentence explaining what the notification is for, acceptance roughly doubles, and the users who accept come back. This is the cheapest retention fix there is, and it is usually one screen and one sentence.
Propose where the first win happens in our onboarding and draft the pre permission screen that asks for push right after it, with one sentence that says what the notification is for, in every language we ship. Estimate the effect on day 3 retention from the users who currently decline. I will ship it in the next release and you measure the cohort after.
Fix two: put what the reviews ask for into onboarding
If 38 reviews ask for a capitals mode, show it on screen two. Onboarding should demonstrate the thing people say they want, not the thing you built first. The Community agent already reads the reviews; the retention fix is to route the most requested feature into the first session.
From the last 200 reviews, list the three features people ask for or praise most. Tell me which of them a new user reaches in the first session and which they do not. Propose an onboarding change that puts the most requested one on screen two, and write the copy for it.
Fix three: win back lapsed trials on day seven, and stop involuntary churn
Send a win back offer to lapsed trials on day seven with the one feature they did not reach. A geography game moved day 30 retention from 17 to 21 percent with the push fix alone, and the win back added paying users at zero acquisition cost. On Google Play, two 2026 settings do part of the job for you: delayed charging retries failed transactions inside the app and cuts involuntary churn by up to 18 percent in Google's own figures, and the account recovery window doubled from 30 to 60 days.
Draft a win back message for trials that lapsed without paying, sent on day seven, that names the one feature they did not reach, in every language we ship, for push and email. Check that delayed charging and the 60 day recovery window are enabled in Play Console and turn them on if not. Report recovered users and recovered revenue per month.
Why retention decides your ad budget
A user who leaves on day 3 never pays back an ad. Every retention point raises the value of every install the organic and paid loops bring. Fix retention before you scale spend, and your payback window gets shorter without changing a single bid, and your rank rises without changing a single keyword.