How we test and rate money-making platforms
There are hundreds of articles ranking the best survey sites. Almost none of them say where their numbers came from. That is the gap this site exists to fill, so it seems only fair to show exactly how we work.
The short version
We sign up with our own details. We run a timer on every session. We count the time spent being disqualified as working time, because you spend it too. We withdraw the money before we publish a rating. And we say when something is not worth your evening.
1. We sign up like a normal user
Every platform we review is one we have joined ourselves, with real details, on a normal account. No press access, no demo logins, no review copies. It also means the screenshots on this site are ours, with personal details redacted. If we have not signed up, we do not publish.
2. We run a timer, and we count the dead time
This is the single biggest difference between our numbers and most others. Every session is timed from opening the platform to closing it, and that includes screening questions, surveys we were disqualified from partway through, time spent looking for something to do, and tasks that failed to credit.
All of it counts as working time, because it is. If you spend twenty minutes being screened out of three surveys and then eight minutes completing one that pays $1.50, you did not earn $11.25 an hour. You earned $3.21 an hour, and any figure that says otherwise is measuring the platform’s convenience rather than your reality.
Most published rankings quote pay per completed task. That is why our ordering often differs from theirs, sometimes by several places. We aim for a minimum of six hours across at least three weeks before publishing a rating, because several platforms improve as their matching learns your profile.
3. We cash out before we rate
A platform that shows a balance has proved nothing. A platform that puts money in your account has proved the only thing that matters. So we withdraw before we publish, and we record the exact amount, the method, how many days it took, and any fee or friction that was not disclosed upfront.
A platform that will not pay is rated on that, not on its promises. If a withdrawal fails we contact support, wait 30 days, follow up once, and if it is still unresolved the platform goes on our list of ones to avoid with the dates and amounts documented.
4. How ratings are calculated
Every score out of five is a weighted average of four sub-scores, never typed in by hand: pay 40%, payout speed 25%, ease of use 20%, support 15%. The weighting reflects what readers actually care about, in order.
Deriving the score rather than assigning it is deliberate. Hand-typed overall ratings drift out of line with their own sub-scores across dozens of reviews, and that inconsistency is what makes a ratings system look invented.
Automatic downgrades
- Unpaid or forfeited withdrawals — capped at 1.5 and moved to the avoid list
- Minimum cashout above $25 — capped at 3.5
- No support response within 14 days — support sub-score of 1
- Card details required at registration for a service that never charges — capped at 2.0
- Earnings that depend on recruiting others — not reviewed at all
5. We re-test every month
Rates change, thresholds change, and platforms quietly get worse. Every platform in a live ranking is re-tested monthly, and every ranked page carries a visible update log showing what changed and when.
6. What we do not do
We do not rate platforms we have not used. There are hundreds we have not got to. They are absent rather than badly rated.
We do not accept payment for coverage, rankings or ratings. We earn affiliate commissions on some sign-ups, and those commissions have no input into the order of any list. On several of our rankings the top-placed platform pays us nothing at all — that is the easiest way to check we mean it.
We do not put affiliate links in our avoid sections. Monetising a warning would destroy the only reason the warning is worth reading.
We do not publish earnings claims we have not personally hit.
Found something wrong?
Our figures are one person’s measured experience, and your results will differ — demographics affect survey matching enormously, and geography affects almost everything. But if you think we have something factually wrong, or a platform has changed since we tested it, tell us. We publish corrections with dates, per our editorial policy. Being corrected is cheaper than being wrong.