Cut survey reversals before you pay out

Survey walls take completes back days or weeks later when the respondent turns out to be a duplicate, a VPN user or a speeder, and by then the points are already cashed out. CashoutGuard links duplicate respondents across accounts, flags fake countries, and shows you which countries and surveys your reversals come from.

• Reversals after the cashout: A survey wall reverses the complete after the user already withdrew, so you pay for the same survey twice. On one GPT site, 64% of the money its survey walls reversed had already been cashed out.
• Duplicate respondents: One person answers surveys from several accounts on the same phone or browser, and the survey wall rejects all of them.
• Fake countries: US and UK surveys pay the most, so users from other countries take them over VPNs and residential proxies, with a phone clock that gives them away.

• Reversals measured, not guessed: report reversed completes by API, or let the postback relay record survey-wall reversal postbacks for you, and see your reversal rate by country, survey, traffic source and signal.
• Duplicate respondents linked: accounts that share a device, browser storage, payout wallet or email pattern are grouped, so a second account is caught on its first survey.
• Fake geography: VPN, proxy and hosting IPs, plus a phone clock or language that does not match the IP country.
• Hold before the cashout: every withdrawal gets a score and readable reasons, so you pay real users fast and hold the rest. Once you report reversals, it also says how much of the cashout can still be reversed, so you pay that part later (what the data shows).
• Slow completes, measured: see how much of each wait between the survey click and the complete your survey walls reversed, and count slow completes with an opt-in rule when your own reversals say they are risky.
• Face ID before the cashout (pilot): one respondent behind several panelist accounts shows up as the same face, before you pay them. How it works.

Survey fraud on GPT sites: speeders, bots and duplicate respondents