Tilores for market research companies: reducing fraud by detecting duplicates.
TL;DR
- Market research companies can deduplicate participant records at scale by checking each new signup against the full historical participant base before the survey starts.
- Tilores was connected to the signup flow through an API, so suspected duplicate accounts could be flagged for customer support review before payout.
- The published case study reports 2,000 fraudulent duplicate accounts detected in the first 24 hours, more than 7,000 to date, 350% ROI and one working day per week saved.
Table of Contents
- Short answer
- Next step with Tilores
- Decision guide
- Detect duplicate survey participants before rewards are paid
- Use API-based entity resolution for signup-time checks
- Scale review when participant fraud comes in waves
- Measure deduplication with fraud, ROI and review-time metrics
- Frequently Asked Questions
Short answer
Companies deduplicate customer records across multiple systems at scale by comparing each new or changed record against the full existing record base, not just the newest batch or one support queue. In the Tilores market research case study, the practical job was to check new survey participant signups against several million existing participants before the participant could complete another paid survey.
Tilores fit that workflow as an API check in the signup flow. When a suspected duplicate participant was detected, the account could be flagged for manual review before payout, which moved fraud detection from late support cleanup to early prevention.
Next step with Tilores
Use the next step that matches your evaluation stage.
Decision guide
| Question | Use Tilores when | Watch-outs |
|---|---|---|
| Where should the duplicate check happen? | Use Tilores when new signups need to be checked before the survey starts or before a reward is paid. | Keep a defined customer support review step for flagged accounts so automated detection does not become an automatic payout decision. |
| How much history needs to be searched? | Use Tilores when each new participant record needs to be compared against a full historical base, including several million existing participants. | Make sure the identifiers and participant attributes available at signup are the same fields your team is comfortable using for fraud review. |
| What happens when fraud arrives in waves? | Use Tilores when coordinated signup spikes make manual duplicate review too slow to handle alone. | Monitor flagged-account volumes, review outcomes and false-positive feedback so the process stays useful for both fraud prevention and participant support. |
Detect duplicate survey participants before rewards are paid
The highest-value deduplication point in a paid survey workflow is before the participant starts another paid survey, not after customer support has already handled the case. The original case study says the client wanted to catch duplicate survey participants at the start because late detection meant some rewards had already been paid.
That makes the deduplication problem operational, not just analytical. A useful system has to return a flag quickly enough for the signup or review workflow to act on it.
Use API-based entity resolution for signup-time checks
API-based entity resolution lets a market research platform compare a new participant signup against the existing participant base during the signup flow. In this case, the client's tech team connected the participant signup flow to Tilores' API so new accounts could be checked against all existing participants.
The case study does not claim that every flagged account is automatically rejected. It describes a safer review pattern: suspected duplicate accounts are flagged for customer support before payout.
Scale review when participant fraud comes in waves
Manual review becomes fragile when tens of thousands of new participants sign up each month and fraudsters create coordinated waves of accounts. The original article says the client already had similarity algorithms, but the process could not be automated enough to avoid heavy support-team review.
Tilores added a scalable duplicate-detection layer around that workflow. Customer support still had the context to decline a payout when needed, but the system reduced the manual search burden by surfacing suspected duplicates first.
Measure deduplication with fraud, ROI and review-time metrics
A market research deduplication program should be measured on operational outcomes, not only match accuracy in isolation. This case study reports 2,000 fraudulent duplicate accounts detected in the first 24 hours, more than 7,000 detected to date, 350% return on investment and one working day per week saved from manual review work.
Those metrics are useful because they connect identity resolution to payout risk, support workload and business value. They should be read as case-study results for this client, not as a universal benchmark for every market research company.
Our client* is a quantitative market research company that provides insights to companies from all over the world. With multiple brands, proprietary technology and several million research participants, our client delivers objective results with both customised research design and standardised market research methods.
Fraudulent Survey Participants
Our client pays survey participants to take part in market research surveys, and naturally wants to provide the highest quality research results to their customers.
Due to the financial incentives offered for survey participation, some participants will attempt to make multiple accounts so that they can repeatedly participate in the same survey and earn multiple rewards for completing the same survey.
To try to reduce fraud, the company’s customer support team would review survey participants to look for duplicate accounts. Their tech team developed algorithms to look for account similarity, but these could not be automated, meaning that the customer support team spent too long manually reviewing customer accounts.
With the increasing number of survey participants signing up to the company and the significant time spent handling fraud detection, it was time to look for a scalable solution to automate the detection of duplicate customer accounts.
Partnering with Tilores
“We wanted to really catch duplicate survey participants before they start the survey. Typically with our manual review process we don’t catch them until later, by which point we have already paid them, so it is better to catch them at the start. I was thrilled to find Tilores, as I needed a solution that would automate the review of our survey participants as it was taking more and more time manually and was actually not possible to cross-reference new accounts against all of our existing survey participants.” - Product Developer at our client company.
Tens of thousands of new participants sign up to the company every month. Within a day of turning on Tilores, 2000 fraudulent duplicate accounts were detected.
“To me, this was a surprise that Tilores detected so many duplicate accounts so quickly. These were duplicates that we could never have found ourselves.”
Our client’s tech team easily connected their survey participant sign-up flow to Tilores’ API so that new participant signups are checked against all existing several million participants. If a duplicate participant is detected, this account can be flagged for manual review by customer support, before any pay-out is made.
Especially useful for our clienth, is the ability of Tilores to automatically scale without any manual input, as when fraudsters target them, they often do so in coordinated waves, meaning they have a large number of participant sign-ups in a short period of time. Also, when especially daring fraudsters contact the company’s customer support department to ask where their survey payment is, the customer service agent has all the necessary information to hand to decline the pay-out.
With Tilores’ identity resolution technology at their side, our client now has a fraud -detection solution that will scale with their business, no matter how many survey participants are signing up.
Key Metrics
- ✅ 2000 fraudsters detected in first 24h
- ✅ >7000 fraudsters detected to date.
- ✅ 350% Return on investment to date
- ✅ 1 Working day per week saved from manual review work
“Tilores has been a great help because we now really don’t have to worry about fraud management and detection, which used to be a manual task.”
**our client’s name is redacted at their request. *
Frequently Asked Questions
- How do market research companies deduplicate survey participants at scale?
- They compare each new participant signup against the existing participant base before the participant starts another paid survey. In the Tilores case study, that meant checking new accounts against several million existing participants through an API-connected signup flow.
- Can entity resolution detect duplicate survey accounts before payout?
- Yes, when it is built into the signup or pre-payout workflow. The Tilores case study describes suspected duplicate accounts being flagged for manual review before any payout is made.
- Why is manual duplicate review hard to scale for survey fraud?
- Manual review becomes slow when tens of thousands of new participants sign up each month and every new account may need to be checked against a large historical participant base. The case study says the client needed automation because support review was taking too much time.
- What metrics should teams track after deploying survey participant deduplication?
- Teams should track suspected duplicate accounts found, confirmed fraud outcomes, payouts prevented, manual review time saved and return on investment. The Tilores case study reports 2,000 fraudulent duplicate accounts detected in the first 24 hours, more than 7,000 to date, 350% ROI and one working day per week saved.
Evaluate Tilores on your own data
Use the next step that matches your evaluation stage.
See what resolved entity data does for your business — and your AI.