What Is a Survey Router? How Survey Walls Find Surveys

What is a survey router?

A survey router is software that sits between people who want to take surveys and the many surveys that are open at any moment. When a user clicks “Start” on a survey wall, the router checks what it knows about them (country, age, gender, profile answers, past attempts), asks a few more questions if needed, and sends them to one survey they are likely to qualify for. If they screen out, a router can offer another match instead of a dead end.

Routers are a normal part of how online research samples are bought and sold. ESOMAR, the global association for the research industry, lists “Do you employ a survey router or any yield management techniques?” as one of the questions in its 37 Questions to Help Buyers of Online Samples, the checklist research buyers use to compare sample providers.

For the user-side comparison of the two kinds of walls, see Survey walls vs offerwalls.

The survey supply chain, from research buyer to your user

A survey your user sees has usually passed through several companies, and each takes part of the price the research buyer pays.

Many survey wall companies play several of these roles at once: they keep user profiles (a panel), buy surveys from marketplaces and direct clients, and run their own router. If you run a rewards site, What is a GPT site? explains the wider business model.

Quotas, prescreeners, LOI and IR

Knowing a few standard terms makes most screen-outs easy to explain.

The definitions come straight from the industry’s own documentation. Cint’s help page on how to find IR calls incidence rate the number of respondents who qualify after demographic and standard targeting in its Exchange (Marketplace) prescreener, and gives the formula completes divided by completes plus non-targetable terms. TapResearch’s survey details guide defines LOI as how long it will take a respondent to fully complete the survey, and says the maximum LOI on its platform is 30 minutes (checked on October 22, 2026).

What happens in one routed survey session

The flow differs by provider, but a typical routed session looks like this:

1. Profile check. The router looks at what it already knows about the user. New users often answer a profile survey first; TapResearch’s publisher FAQ says new users answer its profiler survey, which it then uses to direct them to surveys.
2. Candidate surveys. The router filters open surveys by country, language, device and profile, and drops ones the user already tried or whose quotas are full.
3. Prescreening. For targeting the profile does not cover, the router or marketplace asks a few extra questions before sending the user on.
4. Assignment. The router picks one survey using business rules, such as randomization, priority, or expected conversion. SurveyMonkey’s published answers to ESOMAR’s questions, for example, say its router randomly assigns the panelist to an eligible survey.
5. The survey. The user answers the client survey, which may have its own screener and quality checks.
6. The status. The survey platform sends the user back with a status (complete, screen-out, over quota, quality or security termination), and the wall credits whatever that status pays.

Why users get screened out of surveys

“Why do I keep getting disqualified?” is the most common survey wall complaint, and it rarely means the user did something wrong. Status names vary by platform, but most map to the same few groups. Research Desk’s respondent status codes are a clear example: complete, standard termination, quality termination, security termination and overquota termination.

Cint’s over-quota terminations guide puts the most frustrating one plainly: an over-quota termination occurs when a respondent is disqualified from a survey because quotas have already been filled. That is why a user can qualify in the morning and be turned away in the afternoon.

A few habits also cause screen-outs that look random to users:

• Inconsistent answers. A different age or job in the screener than in the profile can end the survey.
• Guessing what the buyer wants. Claiming to own everything or to be a decision maker is a classic quality flag.
• An incomplete profile. Without profile data, the router guesses, and guesses screen out more.
• Network and device. Location checks compare the IP address with the survey’s country. Cint’s survey protection services page describes a GeoIP check that validates whether a respondent’s IP is local to the survey’s country and language. What is IP reputation? explains why some networks are flagged.

Do survey walls pay for screen-outs?

Some do, and how much varies by provider. Two published examples (checked on October 22, 2026):

• TapResearch’s publisher FAQ describes a reward structure that pays every user in proportion to their time spent answering questions, so users are paid when a survey disqualifies them. Its page on rewarded survey economics adds that disqualifications and abandons happen 1 to 10 times more often than full completes.
• SurveyMonkey’s answers to ESOMAR’s 37 questions say panelists who are disqualified receive $0.10 instead of the full reward, and responses flagged as poor quality receive nothing.

Other walls pay nothing for screen-outs. Find out what your provider does, then tell users. A line such as “You get a small reward even if a survey screens you out” on your earn page sets the right expectation and cuts support tickets.

Quality checks: attention checks, speeding and fingerprints

Research buyers pay for real, honest, unique people, and every layer checks for that before, during and after the survey.

Before the survey

Security checks run when a user enters. Research Defender’s documentation lists digital fingerprinting, bot and click-farm detection, duplication scoring, and a check that detects prior survey attempts before the respondent enters. Cint’s prescreener terminations FAQ describes three kinds of duplication: the same user sent to the same survey again, one person existing as several users in a supplier’s system, and one person entering the same survey from two suppliers. That second one is multiple accounts on your site, and it is your problem to stop; Stop multi-accounting and VPN abuse covers how.

During the survey

• Attention checks. Instructed items such as “select Disagree for this question”, or trap questions about things nobody has done.
• Speeding. Finishing far faster than an honest read allows.
• Straight-lining. The same answer down a whole grid of questions.
• Open-ended checks. Copied, off-topic or machine-written text answers.

After the survey

Buyers review the data, and a complete can be rejected later. TapResearch notes that it expects to be paid for full completes less any deductions after data scrubbing. On your side, that shows up as a reversal of a reward you already credited, which is why survey rewards need the same reversal handling as offers. What is a reversal rate? explains how to track it.

No single check is enough. Pew Research Center’s study Assessing the risks to online polls from bogus respondents found about 4% to 7% bogus respondents in widely used opt-in sources, and that 84% of them passed a trap question and 87% passed a check for answering too quickly. Its 2026 follow-up, No easy fix for bogus respondents in online opt-in polls, concluded that removing bogus cases lowers error on most measures, but that a surefire solution remains elusive. The industry is working on it jointly: the Global Data Quality initiative brings together bodies such as MRS, ESOMAR and the Insights Association to track and reduce fraudulent survey completions. If you also buy survey responses yourself, How to get survey respondents shows how to build these checks into your own surveys.

What publishers should configure on a survey wall

Profiling

Push users to finish the profile survey first and to keep it current. TapResearch’s rewarded survey economics page says first-time users complete a set of profiling questions that is required to be matched to survey opportunities, and that the publisher usually subsidizes this step. A small reward for completing the profile can pay back in fewer screen-outs.

Prefill

If you already know a user’s birth date, gender or postal code, some providers let you pass it so users answer fewer questions. TapResearch’s Survey Feed API overview describes both options: send profile data you collected through its Impanel endpoint, or skip that and let its built-in profiler run. Pass only data the user gave you, only fields the provider documents, and name the provider in your privacy policy. Offerwalls and GDPR covers consent and data sharing.

User messaging

• One sentence under the survey entry: “Surveys look for specific groups, so you may be screened out. Answer honestly and keep your profile up to date.”
• Say what a screen-out pays on your wall, if anything.
• State the rules: one account per person, real location, no VPN or proxy, no rushing. What is a residential proxy? explains why proxy traffic is flagged even when it looks residential.
• Explain that a credited survey can be taken back if the buyer rejects the answers.

Currency, credits and support

Use the same currency and the same value as your other walls; How to set your virtual currency exchange rate shows how. Credit only from the provider’s server-to-server callback, never from the user’s redirect, and handle reversals. In support, separate screen-out complaints (usually expected behavior) from missing completes (worth investigating); Handling missing reward tickets shows what to ask for and where to look.

Measure the right thing

Judge a survey wall by revenue per active user after reversals, not by completes or clicks, and watch how many users finish their profile and come back after their first screen-out. Why users start offers but don’t finish applies the same thinking to offers.

Frequently asked questions