Comparison
Comparing the ways to get 12 testers for 14 days on Google Play: managed services, self-recruitment, bought installs and friends. What each really costs.
“Best closed testing service” is a comparison query, and a comparison only means something once you have decided what is being compared on. A service that is cheapest is not the same as one that is fastest, and neither is the same as the one that leaves you least likely to be refused production access.
For most developers the criteria collapse into four. Does it produce testers with genuine continuous opt-in, because that is the requirement. Does it hold them for the full fourteen days without you chasing. Does it leave you evidence a reviewer can read. And how much of your own calendar does it consume. A provider that wins on three of those is worth paying for even if it loses on price.
The options are more varied than the marketing suggests. You can run the whole thing yourself, recruit a group of people you know, buy installs from a marketplace, or pay a managed service to do it. These differ enormously in what they produce, and the difference shows up in Play Console rather than in the purchase receipt.
Two of them — self-recruitment and paying people you know — are legitimate and can produce a compliant closed test. One of them, buying installs, cannot satisfy the requirement at all, because the requirement is continuous opt-in and a purchase is an install followed by an absence. The managed option sits between the legitimate manual approaches and the illegitimate one, and its value is that it removes the calendar rather than the work you care about.
A managed service supplies the roster, keeps testers opted in for the window, and produces the written record. The reason it is faster is not that it does the work more quickly — it is that the work is done by people who are not chasing strangers on your behalf for a fortnight. Retention is the hard part, and it is a solved problem for a standing pool and an unsolved one for a personal appeal.
The cost is a flat fee per app, with nothing recurring to cancel. What you give up is control over who is testing and when; you are buying a timeline rather than choosing a roster. That trade is usually right when a launch date is fixed, and usually wrong when you have months and enjoy the recruiting.
Recruiting your own roster is entirely legitimate and it costs nothing but your own time. It is also the option most people abandon, and the reason is always the same: interest converts badly. A forum post brings a burst of sign-ups that decays within days, and the people who joined from a favour do not reopen the app a week later.
The arithmetic is unfriendly. Each churned tester costs you the days you thought you were counting, because a replacement starts their own fourteen days from their own opt-in date. Two or three weeks of recruiting is a normal outcome, and the fourteen days begin only once the roster is genuinely complete and stays that way.
Buying installs from a marketplace is the option that looks cheapest and is not a closed test. The requirement is twelve testers who opted in and kept the app installed for fourteen continuous days. A purchased install is an install followed by an absence, so it contributes nothing to the number that matters — and an install pattern with no usage behind it does not resemble genuine opt-in in the way a reviewer is looking for.
The risk is not just wasted money. Using bought installs to satisfy a stated requirement risks a refused production access application, and a refusal costs more calendar than the recruitment you were trying to avoid. If a provider quotes you a price that is far below everyone else, that gap is usually the answer to your question about what they are selling.
Asking people you know is the fastest possible start and the weakest possible roster. They are not independent testers, they install out of politeness, and they rarely return to the app. The count in Play Console can look correct while the evidence behind it is thin, which is the worst combination: it looks complete and it is not.
Using your own devices and accounts to pad a roster is a more serious version of the same error, because it is not independent at all. If you go this route, treat it as an internal test of the app rather than as the closed test that satisfies the requirement, and budget for recruiting genuinely independent testers afterwards.
A comparison that only ever concludes in its own favour is marketing, and it is worth being direct about the cases where buying testers is a bad fit. If you already have a roster you trust and months to spare, recruiting yourself is cheaper and you get to choose your testers. If you only want a sanity check that the app installs and launches, internal testing does that in minutes and nothing here is needed.
This service also cannot shorten the fourteen days, cannot touch the Play Console review queue, and will not help if your problem is a store listing or a policy declaration rather than a tester count. If the refusal you are working around is about anything other than the roster, read what Google actually asked for before spending anything here.
If you would rather vet a provider than take anyone’s word, there are a few questions whose answers are diagnostic. Ask how the roster is recruited and whether the testers are independent. Ask how retention is achieved without the developer chasing people. Ask what happens to a tester who opts out mid-window. Ask what the deliverable looks like at the end, and whether it is something a reviewer could read.
A provider who answers those four concretely is running a testing process. A provider who answers them in terms of guarantees and refunds is selling certainty rather than testing, and certainty is not what the requirement asks for. Price matters last, once the four answers are in front of you.
It depends on which criterion you are optimising. A managed service wins on speed and retention, recruiting yourself wins on cost and control, and internal testing is better than either for a quick install check. Buying installs is not a closed test at all, because it cannot produce continuous opt-in over fourteen days.
A complete roster within hours, retained for the full window, with a written report behind the production access application.