Sample exchange
Buyers describe an audience, panel owners supply it, and both sides read the same record of what was delivered and what was removed. The rejection log is not a report produced on request — it is how the two sides agree on what actually happened.
How the exchange works:
Incidence is measured against stored profile data from our 250,000-member proprietary panel, then quoted with a price, a date and the sourcing mix behind it.
Identity, device and network checks run through Verisoul before an invitation is sent. Inattention is handled separately from fraud, the way the Insights Association defines it.
Buyer file and supplier statement carry the same removals — the rule that fired, the observed value, the timestamp. Partner supply is labelled rather than blended away.
Panel book
Every provider publishes a market count. This is the other half — the audiences where supply is thin, and the ones we will decline rather than quote. Indicative only while the panel is being built. Ask and we will tell you what is actually reachable this week.
| Audience | US | UK | DE | FR | JP | BR | IN |
|---|---|---|---|---|---|---|---|
| Consumer, nationally representative | fills same week | fills same week | fills same week | one week | one week | fills same week | one week |
| Professionals, any level | fills same week | fills same week | one week | one week | two weeks | one week | one week |
| IT decision-makers | one week | one week | two weeks | two weeks | thin, ask first | thin, ask first | two weeks |
| Budget holders, 500+ employees | two weeks | two weeks | thin, ask first | thin, ask first | not panel work | not panel work | thin, ask first |
| Clinicians and allied health | two weeks | two weeks | thin, ask first | not panel work | not panel work | not panel work | thin, ask first |
Updated 2026-09-14 · Ask about an audience
Real people, real opinions
Most sample is found, not kept. Someone is recruited for a study, answers it, and is gone — so every project starts by discovering who these people are. Our members are profiled before a study exists, which is why incidence can be counted instead of estimated, and why the answer arrives in days rather than after a recruitment drive.
Profiling is asked a question or two at a time across weeks, never fifty at signup. People answer carefully when they are not rushing to reach the paid part, and asking the same attribute in different forms months apart is what makes a contradiction visible.
Being ready is only half of it. The other half is being real, and that is not something a panel should grade itself on. Every signup and study entry runs through Verisoul, an independent identity and fraud platform — not our own score marking our own homework.
Three shapes of work make up most of what we field. Each fails in a different way, so each is quoted and monitored differently.
Low-incidence B2B
Infrastructure buyers with signature authority at companies over 500 people. Perhaps four in a hundred panellists qualify.
How it fails: not cost, but screener burn. An optimistic quote exhausts the qualifying population before quota fills, and the next wave has nowhere left to go.
Interlocked consumer
National representation on age, gender and region is easy to hit on each dimension separately and easy to miss on all three at once.
How it fails: weeks later, in analysis. A subgroup turns out to rest on nineteen people who all arrived within two hours from one source.
Repeat rated tasks
Rated work needs the same people back week after week, with reliability tracked across sessions rather than judged inside any one of them.
How it fails: quietly. Raters not paid for sessions that go nowhere stop reading carefully, and the drift shows up as noise nobody can attribute.
The deliverable nobody else sends
Because a supplier who never reports removals is either not looking or not telling, and both are worse than a number that occasionally looks bad.
The log lists every respondent we took out of your file, the rule that removed them, the value that triggered it and when it fired. You can reconcile it line by line against your own counts, which is the only way to know whether quality meant anything on a given project.
The same applies to sourcing. Blended supply is normal in this industry and there is nothing wrong with it. Hiding the blend is what produces the surprises — a subgroup that will not replicate, a result that moves when you repeat the study elsewhere.
Different objectives put pressure on different parts of a sample. These are the ones we are asked for most, and what each one actually demands.
A one-line definition is enough to get a real answer. If it is not reachable from panel supply, we will say so rather than quote it and miss quota in week three. Cash out from $5 if you are joining as a participant instead.
Everything a buyer would normally have to ask for in a call is already on this site.
