Pre-launch report and Android vitals before production
· 6 min read
Somewhere in your closed test, Play Console starts showing two things you didn't set up: a pre-launch report and Android vitals. Both talk about crashes. Neither tells you plainly what to do before you apply for production. This post explains what each one is, which numbers Google publishes, and what is worth your time with 12 to 20 testers.
I read each Google page named below on 10 November 2026. Where Google doesn't say something, I say so instead of guessing.
The pre-launch report: Google tests your app for you
Google's page on pre-launch reports says they are "automatically generated when you upload an app bundle or APK, subject to capacity within our device lab." They also run when you save a release to the production track. To get a new report, you publish another app bundle.
The setup page says a crawler will "automatically launch and crawl your app for several minutes", typing, tapping and swiping. The report is powered by Firebase Test Lab. Its checks cover four areas: stability, Android compatibility, performance and accessibility. The devices run Android 9 and above and include phones, tablets, Wear OS and Chromebooks.
Google also lists limits that matter for a small app:
- Apps without a main launch activity can't be tested.
- Test devices can't make purchases, so subscription and in-app product flows are limited.
- Apps that aren't in open testing or production may fail licensing checks and appear unlicensed.
Give the crawler a login
If your app needs an account, the crawler gets stuck at the sign-in screen and tests nothing behind it. In Play Console go to Test and release, then Testing, then Pre-launch report, then Settings. Under "Test-account credentials" choose "Provide credentials", enter a username and password and save. Use a test account, not a real one. Google says the credentials work only in apps that use standard Android widgets, not WebView-based sign-in. If your app supports Sign-in with Google, Google logs in automatically.
Read the results, and fix in this order
The Overview tab shows Stability, Performance and Accessibility. The Details tab adds stack traces and screenshots. Google's page Understand your pre-launch report sorts findings into three levels:
- Errors (red): crashes, ANRs, defective libraries, and use of unsupported APIs that are already restricted.
- Warnings (yellow): slow startup and load times, sign-in or crawl issues, memory issues, and unsupported APIs not yet restricted.
- Minor issues: missing content labels, color contrast, small touch targets.
Google doesn't give a fix order. My suggestion is to follow its own levels: errors first, then sign-in or crawl warnings, because a crawler that can't get past your login proves little. Minor accessibility issues are cheap and worth fixing, but they come last.
Two honest notes from the same page. Crashes found in a pre-launch report "don't affect your crash statistics", so they won't show up in Android vitals. And "Google can't guarantee that tests will identify all issues."
Android vitals: what real users hit
Android vitals covers "stability, performance, battery usage, and more", based on devices of users who chose to share usage and diagnostics data. Google's developer page lists several core vitals. Two of them matter first for a new app: user-perceived crash rate and user-perceived ANR rate. An ANR ("Application Not Responding") means the app froze and stopped reacting.
The thresholds, from the crash page and the ANR page:
- User-perceived crash, overall: "At least 1.09% of daily active users experience a user-perceived crash, across all device models."
- User-perceived crash, per device: "At least 8% of daily active users experience a user-perceived crash, for a single device model."
- User-perceived ANR, overall: "At least 0.47% of daily active users experience a user-perceived ANR, across all device models."
- User-perceived ANR, per device: "At least 8% of daily users experience a user-perceived ANR, for a single device model."
What happens when you pass them, in Google's words: "If your app exceeds the overall bad behavior threshold, it is likely to be less discoverable on all devices. If your app exceeds the per-device bad behavior threshold on some devices, it is likely to be less discoverable on those devices, and a warning may be shown on your store listing." Google also says Play generally uses the last 28 days of data.
Will a closed test even show data?
Google gives no minimum install count. It says "we only show data if we have enough to make anonymous reports" and that an empty view means "your app doesn't have enough data points within the specified filters." So with a dozen testers you may see nothing. Vitals need real installs from users who share diagnostics. If you do see numbers, one crash among a few users is a big percentage, so read the stack traces, not only the rate.
What to do in a closed test
- Add crash reporting. Android vitals can stay empty, and it only counts users who opted in. Firebase Crashlytics gives you reports for crashes, non-fatal errors and ANRs. Setup is a Gradle plugin and an SDK dependency, then a forced test crash to confirm reports arrive. Note that Crashlytics counts per session and Android vitals counts per daily user, so the percentages won't match. If you add Firebase, update your Data safety form answers too.
- Read the pre-launch report after every upload. Fix the red items first.
- Reproduce each crash from the stack trace and the tester's steps, fix it, and note it in your log.
- Ship a new version code to the closed track. Each one triggers a new pre-launch report.
Write the crashes you fixed and the version codes in your log. They are honest answers for the application, which is covered in Google Play production access. The 14-day rule itself is in Google Play 12 testers for 14 days.
Where TestSwap fits, and where it doesn't
TestSwap is a free site where app makers review each other's apps. You test someone's app, they test yours. Testers use real phones, so they can tell you in writing about crashes and screens that get stuck, with the steps that led there. A clear task helps: how to write a test task.
The pre-launch report covers devices you don't own. A TestSwap reviewer covers a human using the app the way people do. Use both. Neither guarantees approval, and Google decides that. A TestSwap review is a form on our site, not a Play opt-in, so it doesn't count toward your 12 testers unless that person also joins your closed test.
Checklist
- Test-account credentials are set in the pre-launch report settings, if your app has a login.
- Red items in the latest pre-launch report are fixed or understood.
- Crash reporting is in the app, and a test crash showed up.
- Every reported crash has a fix in a new version code on the closed track.
- You know the thresholds: 1.09% crash and 0.47% ANR overall, 8% per device model.
Try it
Post your app on TestSwap with a task that names the screens you worry about. You get 3 credits to start, and reviewing someone else's app first is the quickest way to see how it works. Tell me what's missing or broken (@dizid on X).