We build the plumbing of online sample, and we are honest about where it leaks
ResearchDart routes respondents between suppliers and surveys, screens them before the questionnaire, and keeps the record both sides reconcile against. This page is about how we think, not what we sell.
Why ResearchDart exists
Online research runs on redirects. A respondent leaves a panel, enters a survey, and comes back. At every handoff someone needs to know who the person is, whether they should be there, and how the session ended.
For most of the industry's history those handoffs were treated as setup work: an entry link pasted into an email, four endlinks added the hour before launch, a spreadsheet of IDs sent back weeks later with "quality" typed next to half of them. That was tolerable when the main quality problem was bored humans.
It is not tolerable now. Software that reads questions better than most respondents can be pointed at surveys for a few cents a complete. The checks that matter most have moved out of the questionnaire and into the handoffs, which is exactly the part nobody owned.
So we built the handoffs as a product: signed links, screening before the first question, one outcome per session, and reversals that use the same removal codes the industry associations published.
What we believe
- Stop fraud where it is cheapest
- A respondent removed at the entry link costs nothing. The same respondent removed after cleaning costs a complete, analyst time and an argument with a supplier.
- Measure before you block
- A blocking rule nobody has tested against their own data is a guess. Monitor mode exists because we would not trust our own defaults on your traffic without evidence.
- Reasons beat verdicts
- "Bad quality" teaches a supplier nothing. A GDQ code and a one-line note tells them which source to fix.
- Respondents are giving up their time
- They deserve to be turned away early, told why in plain language, and returned to their panel quickly.
- Survey answers are not ours
- We handle participant IDs, timings and outcomes. Answers stay in the buyer's survey platform.
What we will not claim
A lot of copy in this industry is written to be quoted in a procurement deck. We would rather be quoted accurately.
- A panel size
- We do not own respondents, so we do not publish a reach number.
- A fraud catch rate
- Any single number depends on the traffic, the study and the provider. Measure it on your own projects in monitor mode.
- That screening makes cleaning unnecessary
- It does not. Attention, honesty and open-end quality still have to be checked in the data.
- Customer logos or testimonials we cannot back up
- When customers agree to be named, they will be. Until then, the site shows none.
Who ResearchDart is not for
Saying no early saves both of us time.
- Single-panel trackers
- If one panel already fills your tracker consistently, adding a routing layer adds duplicate risk and reconciliation for little gain.
- Teams that do not clean data
- Screening narrows the problem. If nobody reviews the data afterwards, the part screening cannot see goes straight into your results.
- Anyone looking for traffic without scrutiny
- Every session is recorded with a reason, and suppliers and buyers can both see it. That is the point.
If this sounds like how you want to run fieldwork, talk to us
Tell us what you buy or supply, where it breaks today, and what you have already tried.