Keeping a record of what each panel actually paid

The number on the tile isn’t the number you earn

Every survey panel shows you a headline figure before you click in: an estimated time and a reward. That figure is a forecast, not a receipt. It comes from whoever bought the study telling the panel roughly how long their questionnaire runs and roughly what fraction of people will qualify. Panels pass that forecast straight to you as if it were a promise. It isn’t one, and the mechanism behind why it isn’t is worth understanding before you decide whether any given panel is worth your time.

This isn’t a pitch for survey work as a way to replace an income. It won’t be, for almost anyone. What it can be is a low-effort way to get paid for spare minutes, provided you know which panels on your list are actually giving you a fair rate for those minutes and which ones are quietly wasting them. You can’t tell the difference by reading marketing copy. You can only tell by keeping a record.

Where the advertised numbers actually come from

Market research clients pay panels a fee per completed interview, usually called CPI (cost per interview). That fee is set based on how hard the study is to fill: a niche B2B audience costs more per head than “anyone who drinks coffee.” The panel takes a cut and passes the rest to you as the reward.

The time estimate works the same way. The client (or the panel’s own algorithm, if the client didn’t supply one) estimates length of interview, or LOI, from a handful of test respondents before the study goes live. That’s a small sample making a claim about how long a much bigger, much more varied set of real respondents will take. Some people read every word. Some people speed through it. The number on the tile is an average dressed up as a fact.

Neither of those inputs is wrong on purpose. They’re just estimates built before the study has any real data behind it, and they stay fixed on your screen even as the actual respondent pool proves them off.

Screen-outs are the business model, not a bug

Every study needs a specific mix of people: a quota by age, income, product use, whatever the client is researching. Panels can’t know who qualifies until they ask, so the questionnaire opens with screening questions, and most people who start a study don’t match the quota. That’s expected and priced in from the client’s side. It is not priced in for you the same way.

A screen-out after two minutes of screening questions costs you two minutes for little or no reward, while a full completion pays the same reward for potentially fifteen or twenty minutes of work. If you only ever look at the advertised per-survey reward, you’re comparing offers that have wildly different real costs attached to them, because the screen-out rate isn’t shown to you and varies a lot by panel, by category, and by how narrowly your own profile matches what’s currently in demand.

This is also why the same panel can feel great for one person and terrible for another. If your demographic profile matches what buyers are currently sampling for, you get routed into more full completions. If it doesn’t, you eat more screen-outs. That’s the panel’s matching algorithm doing its job, not evidence that the panel is rigged against you specifically.

What you’re actually trying to measure

The number that matters is pay per minute actually spent, averaged across everything that panel sent you, completions and partials and screen-outs alike. That’s the only number that tells you whether a panel is worth opening again.

You get that number by dividing total payout from a panel over a stretch of time by total time spent on that panel over the same stretch, including every screen-out. Nobody but you has both halves of that equation, because the panel only sees what it paid you, not how long you sat there getting bounced.

A log you can actually keep

You don’t need special software. A spreadsheet with these columns does the job. Log the panel name and the date, then the advertised time and reward the tile showed before you clicked, and the actual time you spent, either from a stopwatch or by noting the clock time before and after. Record the outcome as full completion, partial/screen-out, or disqualified, the actual payout for that entry including zero, and the redemption status showing whether it’s still sitting as points or has actually been cashed out.

After a couple of weeks, add two summary columns per panel: total minutes logged and total payout earned. Divide one by the other. That’s your real rate for that panel, not the one on the tile.

Do this across three or four panels at once rather than one at a time. Panels rotate what’s available to you based on demand, so a panel that looks dead this week might have a run of good studies next week. A short window on a single panel tells you less than a longer window compared across several.

What the pattern tells you over time

Once you’ve got a few weeks of entries, look for a few things:

Drift between advertised and actual time. If a panel consistently runs longer than its own estimate, that’s a structural pattern in how that panel’s clients write their LOI estimates, not a one-off bad study.

Screen-out rate by category. If health studies screen you out constantly but shopping studies rarely do, that tells you something about how your profile matches current demand, and you can spend your limited minutes on the categories more likely to convert.

Payout that never converts to points, or points that never convert to cash. Some panels credit you promptly. Others sit on a completion for review before crediting it, and a share of those reviews resolve as “does not qualify” well after the fact. If a panel’s log shows completions that later got reversed, that’s worth tracking as its own line, separate from screen-outs.

Redemption friction. A payout that’s stuck above a threshold you rarely reach, or that only redeems as a gift card you don’t want, isn’t really available to you yet. Log redemption status separately from earned status so you don’t overstate what a panel actually paid out versus what it merely owes you.

None of this requires guessing at what any panel “really” pays. It requires writing down what happened to you, specifically, on the studies you were actually offered. Two people on the same panel can get genuinely different results because they’re being matched to different studies, so there’s no single true number to look up. The log is the only way to get your own answer.

Account health is a side effect of the same mechanism

If a panel starts sending you fewer studies over time, people usually assume a ban. The common explanation is the same matching mechanism working in reverse. If your recent history skews toward screen-outs or fast disqualifications, some panels quietly reduce how often they route studies to you, because you’re statistically less likely to convert and clients don’t want to pay for a screening funnel that doesn’t fill. Answering consistently and taking your time on the questions that determine eligibility keeps you in the pool that gets routed toward full completions more often. There’s no trick that makes an account immune to this, and no setting that guarantees you’ll keep getting sent studies. It’s a statistical routing decision the panel remakes about you continuously, based on your own history on that panel.

The honest takeaway

Most panels pay unevenly, and the advertised rate is a forecast built on someone else’s small sample, not a quote for your time. The only way to know if a specific panel is worth keeping in your rotation is to track your own actual minutes against your own actual payout, over enough studies that a couple of bad screen-outs don’t skew the picture. Keep the log for a few weeks, cut the panels that come out worst, and keep going with the ones that don’t.

For more on how panels actually work, from screen-outs to account health to getting paid, head back to the Survey Farm home page.