Screening that finds your people fast.

The cheapest screening happens before anyone clicks your link. Panel platforms hold rich prescreening data — demographics, languages, employment, devices — and filtering there costs nothing and wastes nobody's time. Use it for every criterion the panel can express.

Your own screener exists for the criteria the panel cannot know: behavior (“grocery shopped online in the past month”), ownership, experience with a specific product, a clinical population. Put those questions on the first page, route ineligible answers to the screen-out immediately, and only then let the real survey begin. Screening on page one instead of page five is the single biggest cost lever you control.

The moment payment enters, a screener stops being a measurement and becomes a negotiation with strangers who want the job. The rule: ask so the qualifying answer is invisible.

The second version reads like a demographic question. The person who qualifies doesn't know which option mattered, and neither does the person who doesn't. For higher-stakes eligibility, verify with a follow-up only a genuine member of the group answers easily — an EV owner knows roughly where they charge; a faker improvises.

Resist the mirror-image mistake, too: screeners so long they become the study. Every screening question is unpaid work for the majority who won't qualify. Two or three questions is a screener; ten is a survey with a trapdoor.

Quotas cap segments so your sample arrives shaped the way your analysis needs it — 100 per condition, 50/50 gender, no more than 40% from any one age band. Two mechanics matter more than the rest:

Enforcement must be server-side and race-safe. Panels deliver participants in bursts; when the last slot of a cell has thirty people mid-survey, client-side counting oversubscribes it every time. The count-and-close has to happen atomically at the database, not in the browser.

Full cells should turn people away early.The respectful version of “quota full” happens at the first qualifying answer, not after the last page. In Plumeform, quotas are conditions with a limit; full segments screen out at the page boundary where they qualify, the enforcement runs inside the database under a lock, and screened rows never count against your response cap or monthly limit.

A screened-out participant did nothing wrong — they answered your questions honestly and lost the lottery. Two obligations follow. Resolve their submission properly:give them the panel's screen-out exit (a separate completion code and redirect) so their status resolves instead of hanging at awaiting review; never reject someone for failing to be eligible. And write the exit like you mean it— “this study needs a specific mix of participants and this one isn't a match; your time is appreciated” costs nothing and is the difference between a participant who returns for your next study and one who leaves you a one-star review on the panel forum.