On October 6, 2026 an email arrived in a developer's Play Console inbox: "Congratulations! Your app has been granted Google Play production access." The app was Nada, and it was the first app that went through a paid campaign on PeerPlay.
This is the full case study, with the numbers taken straight from the campaign report. I'm writing it down because one real paid campaign taught me more about closed testing than a long list of free ones, and because the same lessons apply to anyone who is about to start their own 14 days.
What Nada Is and What the Developer Bought
Nada is a women's health app built around cycle and period tracking. It has logs with notes, schedules and reminders, and a step where the user takes a photo with the camera. The developer already had the app on PeerPlay's free reciprocal tier with a few testers on it.
They upgraded to Starter Pro, which costs $24.99 one-time and gives 15 verified testers with no reciprocal testing needed. The testers on that plan were approved by hand and test on physical devices. The upgrade kept the testers the developer already had, so nobody had to opt in twice.
What I Had to Fix Before Taking the Money
Having a paying customer made me read my own campaign code with different eyes, and it was not ready. A paid campaign had no fixed end date. Every tester ran their own 14-day round, and a restart button could have paid every tester for another round the customer never bought. The end-of-round tester bonus could also be collected by someone who skipped most of the days.
While Nada's campaign was running I rebuilt that part. A paid campaign now has one calendar for everyone, the bonus needs at least 12 qualified days out of 14, and the server pays it instead of the phone. Your first customer is the most honest code review you will get.
The 14 Days in Numbers
The campaign ran from September 16 to October 6, 2026. In total 18 testers joined, 11 of them completed the full 14 days and 3 quit along the way. The daily log has 217 tester-days, and every one of them qualified.
Day one was slow, with only 7 testers active. From the second day the count stayed between 12 and 15 every day, and it never dropped below 12 between September 17 and October 1. The devices covered Android 9 through Android 17 on Samsung, Xiaomi, OPPO, Google Pixel, Nothing, HONOR, Motorola and Infinix phones.
Most daily sessions in the log are under a minute and still count as qualified. What carried the test was not long sessions, it was the same people opening the app every single day.
What the Testers Actually Found
I expected testers to be a number Google wants to see. They turned out to be QA. The report lists 14 feedback items, and the developer marked 8 of them as solved during the test.
Some of the bugs: turning off a schedule with its switch did nothing, and deleting a schedule did not update the screen until you left and came back. Cycle tracking stopped somewhere around the 19th cycle. A new period entry did not properly close the previous one. The quotation marks in log notes were reversed, and the privacy statement had extra spaces in a sentence. One tester reported that the app did not launch on their phone.
There were feature ideas too: a dark mode for use at night, the front camera (or a switch between cameras) for the photo step so people can see the preview, and health tips during periods. None of this would have shown up on the developer's own phone.
Five Lessons for Your Own Closed Test
Plan for a slow first day. Only 7 testers were active on day one, so count your 14 days from when you really have 12 people, not from when you sent the invites.
Get more testers than the minimum. 3 of 18 testers quit, and the extra slots are what kept the daily count at 12 or more. With exactly 12 people, one dropout means trouble.
Daily presence beats session length. Short daily sessions qualified; skipped days are what hurt.
Treat testers as QA, not a headcount. Read every piece of feedback and reply. The version of Nada that went to production was better than the one that started the test.
Keep the test running until Google answers. Google doesn't publish exactly what it checks during the production review, so leaving the closed test alone until the approval email arrives is the safe choice.
Top comments (0)