The Evolutionary Advace of Synthetic Respondents

How Mimetic Are ‘Synthetics’?

The promise of synthetic respondents holds enormous appeal for the insights industry. If well-executed, it implies an infinite supply of tireless survey task rabbits, able to stand in for reluctant or scarce, sometimes unreliable, real-life consumers. With the explosion of interest in AI over the past several years, an array of synthetic data vendors is rapidly emerging, making use of different methodologies to generate respondents. The potential for cost and time savings is impressive, but there are serious risks associated with trusting synthetic respondents until we have a clear grasp of the data quality implications and the tradeoffs we might be making. Gaining this understanding is made challenging by the fact that current methods of synthetic sample generation vary both in their general ability to mimic human survey-takers and the specific claims they can legitimately make. Rigorous vetting will be critical to identify usable sources of synthetic respondents, understand how well they can perform at their best, and fully assess their limitations.

Many stakeholders from different corners of the insights industry – commercial customers and research practitioners, traditional panel vendors, and academics – have all joined the conversation, some with theoretical critiques, some with empirical research on synthetic data outcomes. Most of the empirical assessments may be missing the mark. This article explains why.