Here's something I've watched play out again and again: a survey taker does everything the forums say — clears cookies, turns on a VPN, even runs incognito — and still gets flagged as a duplicate on the second account. The IP was never the problem.

The short answer: the moment you open a survey, the panel's fraud tools record dozens of browser signals — canvas, WebGL, fonts, screen, timezone — and hash them into a device ID. Industry-standard tools like Imperium's RelevantID, used by major research firms and built into platforms like Qualtrics, then check that ID against every previous respondent. Same device twice = duplicate, instantly.

Why nobody explains this: almost no competitor content covers panel fingerprinting from the taker's side. Panels don't advertise their stack, and most "survey tips" blogs are still telling you to clear cookies — which hasn't mattered since roughly 2015.

What this means for you: if you run more than one survey identity from one browser, the panel already knows. The fix isn't hiding — it's giving each identity its own genuinely separate device fingerprint.

What RelevantID actually does

RelevantID is Imperium's digital fingerprinting product, and it's about as close to an industry standard as this space gets. Here's the mechanics, pieced together from their public documentation and platform integrations:

  • Passive collection. When the survey page loads, a script collects browser and machine attributes — no permission prompt, no visible step. You never know it ran.
  • Fingerprint generation. Those attributes are combined into a persistent identifier for the device. It's more reliable than a cookie because you can't clear it, and it survives incognito mode.
  • Duplicate screening. The fingerprint is checked against the panel's history (and, in shared-database setups, against other panels' histories). A match on a survey you've already taken — or on an account the panel has linked to you — triggers disqualification or a silent flag.
  • Fraud scoring. Newer versions layer in behavioral signals and machine-learning fraud scores, so it's not just "seen before / not seen" — it's a risk profile.

Qualtrics bakes RelevantID directly into its survey security settings, and Forsta's survey platform documents the integration too — this isn't exotic tech, it's the default plumbing of professional panels.

The signals that make up your fingerprint

Signal What it reveals Can you change it?
Canvas rendering How your GPU draws — highly unique No (without tooling)
WebGL renderer Exact GPU/driver string No (without tooling)
Installed fonts OS + software footprint Barely
Screen + color depth Hardware profile No
Timezone / locale Location claim to cross-check Yes — and it must match your IP
Audio fingerprint Audio stack signature No (without tooling)

Individually these are weak. Combined, they're effectively unique — the EFF's early Panopticlick research found the vast majority of browsers uniquely identifiable from a handful of attributes, and commercial tools have only gotten better since.

What "survey validation" tools add on top

Beyond fingerprinting, panels run validation layers that score the whole session: IP reputation, email risk, phone risk, and behavioral anomalies, returned as a risk score rather than a yes/no. The exact vendor mix varies by panel — the point for you is architectural: fingerprinting answers "have I seen this device?", validation answers "should I trust this session?". You need clean answers to both.

Why the usual advice fails

  • "Clear your cookies" — fingerprinting doesn't use cookies. This does nothing.
  • "Use incognito" — same device, same fingerprint. Incognito changes storage, not signals.
  • "Use a VPN" — changes the IP, keeps the fingerprint, and adds a flagged network. Often worse.
  • "Use a different browser" — helps slightly (different fingerprint), but same machine leaks similar hardware signals, and you still share everything else.

The fix: one fingerprint per identity

This is where I stopped thinking about hiding and started thinking about separation. The only structurally sound answer is isolation: each survey identity gets its own browser profile with its own complete fingerprint, cookies, and storage — so the panel genuinely sees a different device. That's what anti-detect browsers do at the profile level: not "spoofing" one fingerprint into looking different each time (which advanced tools flag as inconsistent), but maintaining stable, distinct fingerprints per profile.

The honest boundary

Fingerprint isolation handles the device layer. It does not — and cannot — handle panels that verify the human: photo ID plus facial recognition. One person, one account there, full stop. Knowing which layer you're actually facing is half the game; the other half is not fighting layers you can't win.

FAQ

Is RelevantID on every survey panel?
Not every one, but it's the most widely deployed fingerprinting tool in market research, and most serious panels run something equivalent. Assume you're being fingerprinted.

Can I test my own fingerprint?
Yes — public fingerprint testers show you what your browser leaks. Useful for understanding the problem; they don't fix it.

Does a new laptop solve it?
Temporarily, for one identity. It's a $1,000 solution to a problem profiles solve for free — and it doesn't scale past two.

Do panels share fingerprint data with each other?
Some do through shared fraud databases and vendor networks. Assume a flag on one serious panel can follow you.