The future of market research is data triangulation

Article by Abigail Stuart, Founding Partner at Day One Strategy.

The Oxford English Dictionary has defined market research as “the work of collecting information about what people buy and why,” or, in simpler terms, “the gathering and studying of data relating to consumer preferences, purchasing power, etc.”.

For decades, this has generally been interpreted to involve asking people questions by facilitating focus groups, conducting in-depth interviews and launching surveys. That is, collecting stated opinions and accepting them as fact.

We’ve also had social and digital behavioural data for years now. This data offers us real-time, unfiltered views of how people actually behave in real-life. And more recently, AI-generated insight has entered the mix, giving us the power to analyse, simulate, and synthesise data at speed and scale. These sources don’t replace traditional market research, though they do challenge its monopoly. The market research industry’s response to AI-insights has been painfully defensive. As if traditional market research is infallible.

But the truth is that there is no one source of data that is perfect. Not even traditional research. Every method has blind spots, biases, and trade-offs. So the future won’t be about deciding between data sets – it will be about triangulating.

Data Source: Advantages versus Disadvantages

Rather than relying on either traditional research, social data, or AI-generated insight, what if the future is about combining them all? Think of an approach that combines the richness of the depth interview or survey, the here-and-now of online behaviour, and the predictive power of AI. Not as competing sources, but as complementary views, each compensating for the blind spots of the other. Done right, this would offer something the industry has been promising but failing to deliver: a deeper, predictive, and more precise understanding of behaviour.

New Paradigm for Concept Testing: Smarter, Faster, More Iterative

Take something as familiar as concept testing. Traditionally, it’s a linear, sequential market research process: write concepts, test with respondents, report on results. But with an integrated data approach, it’s far more fluid and insightful.

Here’s a snapshot of what this looks like:

1. Start with social listening and online behaviour analysis to develop real, data-driven personas of your target audience.
2. Train generative AI tools on social data to create digital personas of your target audience. Use them to pressure test dozens of concepts at their very initial stages and uncover likely objections and enhancements.
3. Validate with traditional research, by taking your filtered and developed concepts into focus groups and surveys with a more targeted, focussed sample to validate and evolve ideas based on real human responses.
4. Run your short-listed concepts through your AI personas again to further refine and finalise. You now have a feedback loop between synthetic and real data.