Your 14 days are done, your testers stayed engaged, and production access is approved. Now comes the part nobody warns you about: getting the build out of the closed track and into the hands of real users without breaking the thing you just spent two weeks proving.
Promotion in Play Console is not a single button. It is a ladder — internal, closed, open, production — and each rung has a job. Skip a rung, or climb it with the wrong build, and you can undo weeks of careful testing in one afternoon.
The ladder, in order
Think of the tracks as widening circles, not as a checklist to rush through.
- Internal testing. Your smallest circle: up to 100 testers, usually your own team. This is where a build goes first, to catch the crash that only happens on a real device and the login flow that only breaks outside the emulator.
- Closed testing. Your 12 testers for 14 days, on the track that counts toward production access. This is the proving ground: real people, real phones, real use over two weeks.
- Open testing. Anyone can join, no invite list. This is your first taste of strangers using the app — people who owe you nothing and will uninstall in seconds if the first screen confuses them.
- Production. The store listing, for everyone. There is no undo button here that puts the app back in a test track.
Every build that reaches production should have climbed every rung below it, in order, with the same version code. A build that jumps from internal straight to production has never been tested by anyone outside your team — and your team is the worst possible test group, because they already know how the app is supposed to work.
Promote the same build — do not rebuild it
This is the mistake that costs people the most time, and it is easy to make: you finish closed testing, feel good about the build, and then make "one small fix" before promoting it to production. A new build means a new version code, and a new version code means the thing you are shipping is not the thing you tested.
- Promote the exact build that passed closed testing. In Play Console, promotion copies the release from one track to the next. Use that. If the build earned your testers' approval, ship that build.
- If you must change something, test the change. A crash fix, a text change, a permission removed — any of it means a new build, and a new build should go through internal testing again at minimum, and closed testing again if the change touches anything a user does. "It is only a one-line fix" is how untested builds reach production.
- Keep your version codes in order. Each new release needs a higher version code than the last, on every track. A build with a lower version code than the one already on a track will not be offered to anyone, and you will spend an afternoon wondering why your update is invisible. Decide your version code scheme before you start promoting — a simple incrementing number, one per release, beats a clever scheme nobody remembers.
- One build, one track at a time. Do not push the same version code to closed and open testing simultaneously while you are still deciding. Promote, wait, check the numbers, then promote again. The ladder works because each step gives you information before the next step gets bigger.
Open testing: your first strangers
If closed testing proves the app works, open testing proves it survives people who do not care about you. That is a different test, and it is worth doing even though it is not required for production access.
- Open the track when the closed build is stable. Not before. Open testers will find the rough edges your 12 testers were too polite to mention, and they will do it in public reviews if the app reaches them broken.
- Watch the same signals, at larger scale. Installs, uninstalls in the first day, crash rate, and the first screens people reach. A feature nobody in closed testing used and nobody in open testing uses is not a hidden gem — it is a feature to fix or remove.
- Expect feedback you did not ask for. Open testers will use the app in ways you never imagined, on devices you never tested, with accounts you never set up. That is the point. Collect the crash reports and the confused reviews, fix the top two or three, and only then think about production.
- You can close open testing again. If open testing surfaces a serious problem, pause the track, fix it, and re-test. An open track is not a launch — it is a bigger test, and you are allowed to stop it.
A short open test — even a few days with a stable build — catches the class of problems that 12 friendly testers never will: the first-run experience for someone who has never seen your app before.
Staged rollouts: production without the cliff edge
Production does not have to mean 100% of users on day one. A staged rollout releases the build to a percentage of users first — 10%, then 25%, then 50%, then everyone — and it is the single best safety net Play Console gives you.
- Start small, on purpose. 10% of your users is enough to show a crash spike or a broken purchase flow in the numbers, and small enough that the damage is contained while you fix it. Starting at 100% because you are confident is not confidence — it is refusing the safety net.
- Decide your steps before you start. A common pattern is 10% for a day, 25% for a day, 50% for a day, then 100%. Write it down. If you decide the steps while watching the numbers, you will either rush them when everything looks fine or freeze when one number wobbles.
- Watch crashes and ANRs at each step, not just installs. A rollout that is installing fine and crashing 5% of sessions is not a successful rollout. Check the crash-free rate and the ANR rate before each increase, and compare them with the previous production build, not with zero. A small rise in crashes on a new build is a signal, not background noise.
- Halt is a feature, not a failure. If the numbers go wrong at 10%, halt the rollout. The 90% who never got the build are not a problem — they are the reason the rollout exists. Fix the build, push a new version code through the tracks, and roll out again. Halting at 10% costs you a day. Discovering the same bug at 100% costs you your rating.
- Do not increase the percentage to "fix" a bad number. If crashes are high at 10%, more users will not dilute the problem — they will multiply it. The only direction a bad rollout goes is back: halt, fix, re-test.
What to check before each promotion
Run this list every time you move a build up a rung, not just the first time.
- Is this the exact version code that passed the previous track?
- Are the crash and ANR rates on this build at or below the previous production build?
- Did the testers on the previous track actually use the new features, or just install the app?
- Is the store listing finished — screenshots from this build, a description that matches what the app does, a privacy policy link that works?
- Are the permissions the app requests the ones it actually uses, with nothing left over from an earlier experiment?
- Do you know what you will do if the numbers go wrong at the next step — halt, and at what threshold?
If any answer is no, the build is not ready to climb. Fix it on the lower track, where the audience is small and forgiving.
The promotion checklist
- Promote the exact build that passed closed testing — same version code, no last-minute rebuild.
- Run a short open test with the stable build and fix what strangers find.
- Start production as a staged rollout: 10%, then 25%, then 50%, then 100%.
- Check crash-free rate and ANRs before every increase, against the previous production build.
- Halt the rollout at the first real signal — do not push through it.
- If you fix anything, the fix is a new build, and a new build climbs the ladder again.
Fourteen days of closed testing earns you the right to launch. The ladder is how you use that right without gambling it on a single release.
The part no tool does for you
Play Console gives you the tracks, the promotion button, and the rollout percentage. What it does not do is find you 12 real people with Android phones and different Google accounts, get each of them opted in on the right account, keep them installed and engaged for 14 continuous days, replace the ones who go quiet, and hand you a final report built for the production access form — so that the build you promote is one that real testing actually proved.
That is exactly what VerifiedTesters handles: 12 real testers on their own Android phones, opted in on your Play Console closed track and engaged for 14 continuous days, with free dropout replacement and a final report built for the production access form.
Climb the ladder with a build that earned it — then let your launch count.
Top comments (0)