The Cost of Bad Data: Why Data Quality is More Important Than Ever

Filed Under: Shopper, Advanced Analytics, Data Quality, Quantitative Research

Published:

In the world of market research, the quality of your data is paramount—it’s the foundation upon which critical business decisions are made. We, at C+R Research, take data quality very seriously, which is why we held a special Emerge Smarter webinar on this very subject.

My colleague, Kiara Frey, joined me to talk about the pervasive issue of bad actors—respondents who provide false or misleading information, either unintentionally or through deliberate fraud – in survey research and their potential to distort data that brands like yours rely on for strategic decision-making. We have a Data Quality team, of which Kiara and I are members, to stay vigilant in the constant battle against these bad actors.

Why Data Quality Matters More Than Ever
The consequences of decisions based on compromised data can be severe, leading to misguided strategies that can damage a brand’s reputation and success. Businesses invest heavily in market research initiatives and the quality of the data is vital. The insights derived from this data often drive decisions worth millions of dollars, shaping brand strategies, targeting efforts, and communication plans. However, the accuracy of this data is constantly under threat from bad actors. At C+R, we understand the critical importance of sample and data quality, having spent decades in the trenches, constantly honing our methods to identify and eliminate these bad actors from our surveys.

The Evolution of Data Quality: Adapting to New Challenges
Data quality issues are not new, but as bad actors grow more sophisticated, the complexity of how to combat them continues to increase. Staying ahead in the data quality battle demands a proactive approach that combines both human oversight and advanced technology.

During the webinar, I provided a historical overview of how data quality measures have evolved from traditional methods like mall intercepts to today’s sophisticated online surveys. Moreover, while the challenges have changed, the core principles of ensuring respondent authenticity and engagement remain the same.

The Sentinel System: A Comprehensive Approach to Data Quality
C+R’s commitment to data quality is embodied in our Sentinel System—a robust, multi-layered framework designed to safeguard the integrity of survey data during every stage of our research projects. The Sentinel System represents a culmination of years of experience and innovation in data quality management. It integrates pre-survey reviews, in-survey monitoring, and post-survey analysis, each layer meticulously designed to filter out fraudulent respondents and ensure that the data collected is both accurate and reliable.

Key components of the Sentinel System include:

  • Pre-Survey Screening: Utilizing advanced analytics to flag potential bad actors before they even enter the survey. This stage includes checks on device location, IP address, and other behavioral indicators that suggest fraudulent intent.
  • In-Survey Monitoring: Continuous evaluation of respondents during the survey, focusing on response patterns, consistency, and timing to detect any anomalies that may indicate dishonesty or disengagement.
  • Post-Survey Analysis: After the survey, additional checks are performed, including reviewing open-ended responses and applying advanced analytics as our final step.

Research on Research: Validating the Sentinel System
To demonstrate the effectiveness of the Sentinel System, we conducted a research-on-research initiative, focused on grocery shopping behaviors. The study analyzed data from:

  • A gen pop sample of primary grocery shoppers, along with several food categories that we knew have a relatively low incidence of weekly purchasers. Knowing that bad actors try to qualify for lower incidence groups to receive a higher incentive, we specifically focused our efforts on understanding the impact on these categories.

As we suspected, the data revealed the significant impact that bad actors can have on research results.

  • For example, in the gen pop sample, through the Sentinel System, we identified fraudulent respondents. If those bad actors were left in the study, we saw that many behaviors were inflated – from weekly category purchases to where they shop and even their attitudes towards sustainability.
  • With the Sentinel System, we removed 19% of the gen pop sample who were fraudulent respondents. The results were even more dramatic when we analyzed one of our low incidence category groups (weekly bread mix purchasers), revealing how a brand’s targeting and messaging strategies can be “off the mark” when a comprehensive data quality system is not in place. It’s hard to believe but in this case, we remove almost half (47%) of the fraudulent weekly bread mix purchasers!

Looking Ahead: The Role of AI in Enhancing Data Quality
As we look to the future, AI is poised to play an increasingly important role in maintaining data quality. While AI offers powerful tools for tasks like evaluating open-ended responses, it also presents new challenges, as fraudsters may use AI to their advantage. C+R is actively exploring how best to integrate AI into our Sentinel System, ensuring that we continue to provide our clients with the highest quality data possible.

Conclusion
In an industry where accuracy is paramount, data quality cannot be left to chance. C+R Research’s commitment to excellence, exemplified by our Sentinel System, ensures that the data you rely on for critical business decisions is as accurate and reliable as possible. As the landscape of market research continues to evolve, we remain dedicated to staying at the forefront of data quality, protecting your investments, and helping you make informed, impactful decisions.

In our next data quality blog, I’ll cover key questions every client should ask their research partners. 

explore featured
Case studies

Hey, get our newsletter

join 5,000+ market research professionals
who “emerge smarter” with our insights