Android onboarding tests answer a narrower question than the iOS version, and pretending otherwise is what produces results nobody can act on. You are not testing whether the flow works. You are testing whether it works on a class of device.
What to put in the public brief
- Android, and the minimum OS version
- The device tier you need, described in plain terms
- Whether they will sideload or join a testing track
- That a screen recording is required
- Roughly how long the run takes
Keep the app name, the build and any credentials private.
What to put in the private instructions
Say how they get the build, step by step
Include the unknown-sources warning if there is one, and say it is expected. A tester who stops at a security prompt has produced a recording of your distribution method rather than your onboarding.
Give an outcome, not a tap sequence
"Get to the point where you have created your first entry." A route tests instruction-following; an outcome tests the app.
Ask for the exact device and Android version
Manufacturer, model and build number. "Samsung, recent" is not a data point you can reproduce against.
Ask them to note anything that felt slow
Perceived performance is a first-run killer on mid-range hardware and it never appears in a crash log.
Proof worth requiring
Ask for
The full screen recording, the device model and Android version, and one line on anything that felt slow or looked wrong rather than broken.
Do not ask for
Logs or a bug report format. You are buying a first-run experience from a normal user; asking for structured triage selects for testers who are not normal users.
When this task type is the wrong tool
If you need reproduction of a known crash on a known device, this is the wrong instrument — that is QA with a specific repro, not first-run research.
If the app is not yet installable outside your own team, nothing here can run. And if you want the same question answered on iPhone, the iOS version of this brief differs in exactly one respect: you can stop worrying about which device it is.