Why survey accounts get banned

The account closure that costs you money is almost never the one you saw coming. It arrives with a balance sitting in it, two days before you would have cashed out.

The four triggers

Duplicate identity signals. One person, several accounts on the same panel. This is the big one, and it is banned in the terms of essentially every platform, because the panel is selling unique respondents and a duplicate corrupts the sample it sold.

One account per household, enforced on the IP. Two genuinely different people in the same flat can trip this, and it is the most unfair item on the list. Some panels allow it if you tell them in advance. Most do not, and finding out afterwards is the usual way people discover the rule exists.

Inconsistent profile answers. Panels compare what you say now against what you said before, including inside a single survey where the same question gets asked twice in different words. Contradictions read as inattention at best and fraud at worst.

Speed-running. Finishing a fifteen minute survey in four minutes, straight-lining every grid question down the same column, or failing the attention checks that are seeded through longer studies. Panels score this, and the score is why you can be removed without any single dramatic event.

How panels detect duplicates

Three signals, and they are not secret.

The IP address you connect from, which groups accounts by household and by network. The device fingerprint, meaning the collection of properties your browser and machine expose, which is stable enough to link two accounts that never share a login. And the payment destination, which is the strongest of the three, because two accounts cashing out to one PayPal address is not circumstantial.

That last one catches more people than the other two combined, because someone runs a second account carefully for months and then sends both balances to the email address they have used since 2014.

Careful everywhere except the one place it counted.

If you want the technical version of how fingerprinting works, we cover that in more depth on antidetectreview.org, which is our own site. What matters here is narrower: these signals exist, panels use them, and the rules they enforce are the rules you agreed to.

I will say the unpopular part plainly. If you are running multiple accounts on one panel, you are breaking that panel’s terms, and no amount of technical care makes that not true. What the detection detail is genuinely useful for is the other case, which is far more common: one honest account that keeps getting flagged because of a shared flat, a shared network, or a phone that changes address every time it reconnects.

Payment holds are not bans

A hold is a review. Your account still works, surveys still route, and the balance is frozen while something is checked. It is usually triggered by a first large cash-out, a change of payment address, or a login from a new country.

A ban is a closure that takes the balance with it.

People conflate the two and panic at the first, which leads to opening a support ticket a day, or worse, creating a second account to “keep earning” while the first is under review. That converts a temporary hold into a genuine duplicate violation, which is the one thing on this page that reliably ends an account, and it happens to people who did nothing wrong until the moment they panicked. The hold would have cleared on its own.

Read the review window in your panel’s own terms rather than a figure someone quoted in a forum. They differ, and the panel’s version is the one that applies to you.

Recovering an account

Support will generally reverse a hold once identity is confirmed, and will restore an account closed by an automated flag when you can show the flag was wrong. A shared household is the case they are most willing to look at, because it is common and verifiable.

They will not restore an account closed for confirmed duplicates, and they will not pay out a forfeited balance. Appealing that outcome mostly wastes the weeks you could have spent building a clean account elsewhere.

So the practical move when something goes wrong is to cash out whatever is available on your other platforms first, then appeal once, in writing, with specifics. Not four times.

What I have not tested is how any specific panel’s appeals process behaves, because the only way to generate that data is to get accounts closed on purpose. The mechanics above come from the panels’ published terms and from what people report consistently, and the recovery rates are not something I can put a number on.