Financial Services/Quantitative/North America

Cleaning fraud out of a high-stakes tracker

<1%bad-data rate, sustained
<1%
Bad-data rate
2-layer
Fraud screening
100%
Open-ends reviewed
NA
Region

The fraud screening is the difference. We finally know the numbers are real before we brief the board.

Consumer Insights Director, Financial Services

The challenge

A financial-services tracker was being polluted by bots, duplicates and inattentive respondents. Rejection rates were high, timelines slipped on re-fielding, and confidence in the data had eroded.

Our approach

We rebuilt the quality stack around the study: digital fingerprinting, deduplication and a suppression list before the survey, plus AI-assisted open-end review, speed and straight-lining checks in-field.

The trust & data-quality layer

Screening runs in two layers — pre-survey to stop fraud at the door, in-survey to catch it mid-stream. Respondents are never told why they were removed, so the methods can't be gamed. Every rejection is logged and reported.

Results & impact

  • Sustained bad-data rate under 1% across waves
  • Far less re-fielding, so timelines held
  • Restored confidence for board-level reporting