What Is Synthetic Data in Market Research?
Synthetic data in a market research context refers to artificially generated data that mimics the statistical properties of real consumer responses. Rather than collecting data from actual human respondents, AI models trained on existing datasets can generate simulated responses that, in theory, reflect the attitudes and behaviours of defined consumer populations.
The Promise
The potential benefits are significant. Synthetic data can be generated at a fraction of the cost and time of real fieldwork. It sidesteps the increasing difficulty of recruiting high-quality survey respondents. It allows researchers to run unlimited simulations without additional data collection. And it can generate data for populations that are difficult or expensive to reach through traditional methods.
The Problems
The problems are equally significant. Synthetic data is only as good as the training data it is built on — if that data contains biases or is outdated, the synthetic outputs will replicate and potentially amplify those biases. More fundamentally, synthetic data cannot capture genuine human reactions to new stimuli — it can only extrapolate from past patterns. Testing a genuinely new concept with synthetic data risks circular reasoning.
The Industry Divide
The market research industry is deeply divided on synthetic data. Some pioneering agencies are incorporating it into their service offerings, presenting it as a complement to traditional fieldwork. Major professional bodies including MRS have issued guidance cautioning against its uncritical adoption and calling for transparency with clients about when synthetic data has been used.
The Bottom Line
Synthetic data is a tool — like any tool, its value depends entirely on how it is used. For exploratory research, hypothesis generation and simulation modelling, it has genuine utility. For final-stage consumer research that will inform major investment decisions, real human data remains irreplaceable. The researchers who understand both the opportunities and the limitations will be best positioned to use synthetic data wisely.