Brands make costly decisions on flawed consumer data, a Fortune commentary argues
A Fortune commentary warns that companies routinely act on unreliable consumer research, because people misreport their own motives and behavior, and that AI-generated synthetic audiences risk spreading the same errors faster.
Brands are spending heavily on consumer research and still acting on information that is often wrong, according to a Fortune commentary published in November 2026. The author, who co-founded a research consultancy with a business partner, argues that the problem is not a lack of data but the unreliability of what customers say about themselves, and that the cost of acting on it is substantial.
The piece cites an estimate from ESOMAR, the global association for market and social research, that the insights industry is worth about $150 billion. Market research spending alone is put at roughly $56 billion. Against that backdrop, the author contends that many decisions built on this research are misguided, leading companies to put money into products and features that customers do not end up using.
To illustrate the gap between stated intent and real behavior, the author points to an app developed by his co-founder called Anyone. Survey respondents predicted strong demand and said they would want the app. In practice, the app reached a download peak of about 25,000, but its active user base fell to around 1,000. Many of the people who had said they would use it never downloaded it at all.
The central claim is that people often misrepresent their own motivations, and that they do so frequently without realizing it. The author offers an example from the watch market: buyers tend to explain their purchases in terms of craftsmanship and quality, while the status signaling that may actually drive the decision goes unmentioned. Respondents, in other words, give answers that sound rational and flattering rather than accurate.
The commentary sets out two practical adjustments for researchers. The first is to ask people what their friends and social circle think, rather than relying only on how individuals judge themselves, since opinions shaped by others can be a more reliable guide to behavior. The second is to move beyond a single snapshot in time and ask about past behavior and expected future behavior, which can reveal patterns that a one-off question misses.
The author also warns about the growing use of synthetic audiences, which are AI simulations of consumers that companies employ to test ideas quickly. Because these models are trained on existing data, he argues, a synthetic audience built on flawed consumer information will produce fast answers that are still wrong. Speed, in that case, amplifies the error rather than correcting it.
Source: fortune.com
8
