Why Reward Apps Ban Accounts, and Where the Balance Goes

Why the Message Never Explains the Reason

The notice is always short: unusual activity, a policy violation, a review of your account. No line item, no date, no screen showing which action was the problem.

The silence has a duller explanation. A detailed reason is also a map. If a program published exactly which pattern crossed which line, the people building fake traffic on purpose would read that page first and adjust around it, and the detection would be worth less the day after it was explained. So the message stays generic for everyone, including the large majority of people who did nothing deliberate at all.

That is the first thing worth accepting before reading further: you will probably never be told what the trigger was. The useful question is not which exact click ended the account, because that answer is not coming. The useful question is what kind of thing the system watches for, because that shape is public, documented by the platforms that run the advertising underneath these apps, and it makes the outcome legible even when the individual case stays sealed.

Who Is Actually Paying You, and What They Are Buying

A reward app is a middle layer. The money that ends up in your balance did not start there. In most of these products the revenue arrives from advertising: an advertiser pays for people to see or interact with something, a network places that inventory inside apps, and the app passes a slice of it to the user as points, credit or cash. Your reward is a share of a sale that happened above you.

Because of that, the platforms sitting between the advertiser and the app defend the advertiser first. Google's AdMob help page on invalid traffic states the position without hedging: "If we observe high levels of invalid traffic on your account, we may suspend or disable the account to protect our advertisers and users." The stated purpose of the action is protection of the buyer, not punishment of the seller.

The Rule That Explains Most of the Surprise Bans

Here is the single most useful fact on this page, and the one that resolves most of the confusion about why an ordinary user can be closed out. Action does not require proof of fraud. Doubt is enough.

The same AdMob page continues: "if we are unable to verify the quality of your traffic, we may limit or disable your ad serving." Read that carefully, because the standard it describes is not guilt. It is unverifiability. The question the system asks is not whether you cheated. It is whether the quality of what you produced can be confirmed. If the answer is no, the response is the same as if the answer had been yes.

This is why people who genuinely did nothing still get caught. A perfectly ordinary pattern can be unverifiable for reasons that have nothing to do with the person: a shared connection at work or in a building, a phone that many people have used, a network route that makes the origin ambiguous, a device configuration that reports itself oddly. None of that is misconduct, and none of it makes the traffic confirmable either.

What Counts as a Signal, Described From the Outside

A signal is any observation that makes traffic harder to confirm. It is not a rule you break, it is a measurement that comes out unusual. Several are well documented as categories, and they are described here as things the system notices, never as things to manage.

More than one account reached from the same device or the same connection is the most common. The practice has a name of its own: Wikipedia's entry on sock puppet accounts defines a sock puppet as "a false online identity used for deceptive purposes", and notes that such accounts may be created to "circumvent restrictions". Naming it as a category is the point, because a system that has a name for something has a detector for it.

Location and identity that do not agree register too: when the connection, the account and the payout method point at different countries, there is no single consistent story to verify.

Interaction shaped unlike a person is a third family: bursts too regular to be a hand, sessions with no variation, activity running through environments where a human presence cannot be established. Google Play's deceptive behaviour policy sets the baseline in this direction, saying it does not allow apps that "attempt to deceive users", which is the developer facing half of the same standard the user meets from below.

How One Device Is Recognised Across Several Accounts

People are often genuinely surprised that a new account, a new email and a new name did not read as a new person. The answer is that the account is not the thing being identified. The machine is.

Wikipedia's article on device fingerprinting describes it plainly: "A device fingerprint or machine fingerprint is information collected about the software and hardware of a remote computing device for the purpose of identification." The article adds that the collected information "is usually assimilated into a brief identifier using a fingerprinting algorithm". A phone, in other words, has a description, and that description travels with it whatever name is typed into the sign up form.

Why the Balance Leaves With the Account

The part that hurts is not the account. It is the number that was sitting inside it. People reasonably feel that whatever the dispute is about future access, the amount already earned was already earned.

The credit did not work that way at any point, though the interface rarely says so. In an advertising funded reward, the app is credited by the network after the fact, and it in turn credits you in advance of that settlement. Until the underlying activity is finalised, what you are looking at is a provisional entry. AdMob describes the consequence for the app side directly: "Due to invalid traffic, you may also see a difference between your estimated and finalized earnings." Estimated and finalised are two different numbers by design, and the balance on your screen is the first one.

So when activity is reclassified, the reversal does not stop politely at the account boundary. The revenue that funded the credit is removed above, and the credit built on it goes with it. That is why closure and a zeroed balance arrive together rather than as two separate decisions, and why appealing them is really one appeal, not two.

What an Appeal Realistically Looks Like

Most programs have some route back: a form inside the closure message, a support address, a help centre with a category for account decisions. The route usually exists. What it produces is not something anyone outside the company can promise, and this page does not.

The person reading your message is not deciding whether you are a good person. They are deciding whether the quality of the account's traffic can now be verified where it could not before, so everything useful you send is evidence in that direction.

Practical shape follows from that. Write once, plainly, with the account identifier and dates. Say what your actual usage looked like. Attach whatever confirms it. Ask what the decision was based on, accepting that the answer may not come. Send it to the channel the program names rather than to a public post, and then wait rather than sending the same message five more times, because a queue that has your case will not move faster for being crowded.

Closure Is Not the Same Thing as a Scam

After a closure, the natural conclusion is that the whole thing was fake and the ban was just the moment the mask came off. Sometimes that is true. Often it is not, and the two situations behave differently enough to tell apart.

A legitimate program closing an account is enforcing a rule for the reasons described above. It is entirely possible for a real program to be right about its policy, wrong about your account, and unreachable about the difference, all at once.

The separate category has its own markers, and the clearest one is direction of payment. The Federal Trade Commission's guidance on job scams states it flatly in its own context: "If someone asks you to pay to get a job, that's a scam." The general pattern behind that line is being asked to pay in order to be paid, and it belongs to a different family of problem than a policy enforcement.

What Changes on the Next Account, Without Any Promise

Nothing in this section is a guarantee of immunity, and no article can offer one, because the standard is verifiability rather than compliance and verifiability partly depends on conditions you do not control.

What can be said is which habits leave a cleaner record. One account per person, consistent with real identity, is the base of it. Personal details, connection and payout method that agree with each other and stay that way. Ordinary use on the device you actually own. Reading the terms once before accumulating anything, since the referral rules and the household rules are where most accidental violations live and they differ between programs.

There is also a strategic point that has nothing to do with detection. A balance that sits for months in a program is exposure, and it stays exposed to whatever the program decides later. Withdrawing at the earliest point the program allows converts a conditional entry into a settled one, which is the only reliable way an amount stops being reversible.

And a wider question is worth asking after a closure like this: whether the effort was matched to the risk. If a program required so much accumulated activity that the balance became meaningful before it became withdrawable, the incident is a good moment to check whether the programs you use are worth the exposure and to reduce the number of places where a provisional number is sitting.