Across the research data collection sector, there has been a clear rise in AI-driven solutions designed to detect survey fraud, keep out bots, and clean data more efficiently. This is an important development. As fraudsters and bots become more sophisticated, the industry’s defences have to evolve too.
At Yonder Data Solutions, data quality is an area of continued investment. Our Data Quality Charter is built around multiple layers of defence and we have recently integrated banking-grade fraud prevention technology, Verisoul, directly into our tech stack.
Technology has an essential role to play, but as the industry builds more sophisticated data-cleaning systems, there is a more fundamental question that needs to be asked around quality.
What level of data quality are we starting with?
In parts of the market research industry, there can be a strong focus on the process of cleaning data, while less attention is paid to the quality of the source.
It is a bit like a restaurant investing in the best chefs, equipment and kitchen systems, while compromising on the quality of the produce. The process matters, but the ingredients still shape the final result.
In market research, those ingredients are the respondents. Not all panels are created equal, and the quality of ther respondent source has always mattered. Today, in the age of AI and automated bots, that statement is truer than ever.
Good data starts with good inputs
At Yonder Data Solutions, our Data Quality Charter is built on a multi-layered Swiss Cheese model of defence. Advanced tech and AI are vital layers in that model, but the first layer and the one everything else depends on, is valuing respondents.
Through our proprietary UK panel, Y Live, that means focusing carefully on the respondent experience:
- Fair reward: Members are paid fairly for their time, above the equivalent of the minimum wage, with an average of £2 million paid out to the panel every year.
- Respect and engagement: Y Live is built around a fair exchange, where members feel their time and opinions are valued.
- Authenticity and representation: Engaged respondents are more likely to provide thoughtful, considered and genuinely human responses, particularly in open-ended questions.
Quality has to be built into the economics
Representation, recruitment quality and respondent experience all need to be treated as part of data quality, not separate from it. This raises an important tension for research buyers.
Most clients express that data quality is a priority. But in procurement conversations, there is still pressure driven by the lowest price per complete.
Paying panellists fairly and maintaining a trusted, proprietary community are crucial to the future of the industry. This will only become more important as primary research data increasingly informs AI-enabled analysis and synthetic modelling. If research starts with poor-quality inputs, it should not be surprising when you more time, money and technology are needed to make the final data usable. Technology is essential, but it should strengthen the quality of the research, not compensate for a weak foundation.
When choosing a data collection partner, it is worth asking: are they only investing in technology to clean the data, or are they also investing in the quality of the respondent source?
Get in touch with the team to discuss how we can support your fieldwork and data collection needs: hello@yonderdatasolutions.com
You may also be interested in
Representation in Focus
EXPLORE CAMPAIGN