At Nepa, data quality is at the core of everything we do. As concerns about poor data quality and fraudulent survey responses grow in the market research industry, we remain dedicated to delivering some of the most reliable and accurate insights available. Faced with challenges from bots, professional survey cheaters, and click farms, we combine advanced technical solutions with robust policies to detect, address, and prevent fraudulent entries. We also prioritise respondent engagement by designing short, engaging surveys with advanced logic, cutting-edge research design, and carefully controlled language.
Dynata Rise Awards 2023
Award winning surveys
At the core of our research lies well-designed, engaging surveys. Crafted by industry experts and programmed by a dedicated team, we use one of the leading survey tools to ensure the highest quality responses. In 2023, Nepa received the Dynata Rise Award, recognising us as one of the best in the industry for achieving respondent engagement that leads to better quality answers. We focus on AI, employing a survey scoring tool that automatically disqualifies respondents based on multiple evaluation factors, ensuring valid, engaged responses.
How?
How Nepa combats survey fraud
We take a holistic approach to data quality by integrating internal processes, respondent-centric survey design, and third-party collaborations to uphold the highest standards. We continuously pilot and evaluate new tools and methodologies to stay ahead of evolving fraudulent activities.
Detecting fraudulent responses
Nepa employs multiple strategies to identify and eliminate fraudulent data, evaluating seven key markers:
- Duplicates: Repeatedly submitted identical or highly similar responses.
- Speeders: Respondents completing surveys too quickly, indicating disengagement.
- Inconsistent repliers: Conflicting answers within a survey suggesting fraudulent activity.
- Straightliners: Uniform responses across all questions.
- Randomisers: Erratic responses that defy logical trends, often linked to professional cheaters.
- Open-enders: Nonsensical answers to open-ended questions, flagged through automated and manual checks.
- Multifactor outliers: Using AI-powered tools to detect bots, fraudsters, and survey farms via passive data points.
Automation and manual review
We use a blend of automated systems and manual checks to counter fraud:
- Automated statistical analysis: Built-in tools flag patterns like unusually fast responses.
- Manual review: Time-intensive but critical for catching subtle fraudulent activity.
- Third-party software: Trusted providers identify bots and cheaters with advanced algorithms.
- CAPTCHA: Ensuring only humans participate at the survey entry point.
Continuous monitoring and supplier partnerships
Strong relationships with data suppliers are key. We demand high-quality standards and continuously monitor and adjust strategies to tackle market-specific fraud, ensuring authentic insights.
YOUR TOP DATA QUESTIONS ANSWERED
Webinar on demand
Account Director Martin Hellgren was joined by Nepa’s Head of Data Procurement, Fredrik Olsson, and Dynata’s VP Sales, Benelux & Nordics, Wilko Rozema to discuss your most pressing questions about data quality, market research challenges, and the evolving insights landscape. They shared valuable perspectives on sourcing high-quality data, navigating industry trends, and optimising research strategies for better decision-making.
If you have any further questions or would like to explore these topics in more detail, feel free to reach out – we’d love to continue the conversation!
FAQ
Data quality FAQs
Has fraud increased in market research?
Yes, since 2022, professional fraudsters – including click farms – have increasingly bypassed security measures to earn survey incentives. Major markets like the US, UK, India, and China were first impacted, with Nordic markets seeing similar issues recently. Nepa combats this with engaging, short surveys, API-based checks, and vigilant monitoring of trend deviations.
What indicates fraudulent shifts in data?
Typical signs include sudden drops in well-known brands, convergence toward the mean for brand awareness, and unusual spikes in unaided responses favouring a single brand.
How do you detect bad or fraudulent data?
Indicators include shifts in subpanel distribution, a rise in desktop responses, fewer brand selections in multi-choice questions, longer survey durations, and abnormal KPI levels within specific subpanels.
Looking ahead
Staying ahead of fraud remains a top priority. By blending in-house expertise and third-party technology, we maintain industry-leading data quality. Our proactive adoption of tools like Dynata’s QualityScore keeps us prepared for future challenges.
