Thinking / Leveraging the wisdom of crowds for audience refinement

Leveraging the wisdom of crowds for audience refinement

Prediction Market Research Case

The Challenge

A longstanding and reputable Fortune 100 financial services company planning a new product introduction asked Marcus Thomas to design and execute an in-market, real-world test to validate viability for a national launch. As a precursor to the test, we recommended research to refine the audience – as initial understanding of the target audience was broad. Specifically, we needed to determine and profile consumers most likely to purchase the product to create increased efficiency and effective audience-targeting for the test.

From the start, we faced two challenges:

  1. The need to deliver results quickly and not delay in-market test schedules.
  2. Client skepticism of traditional research methodologies (e.g., concept tests), as they had experienced gaps between self-reported behaviors (i.e., what consumers say) and actual market performance (i.e., consumer purchase behavior) for previous new product launches.

The solution – Prediction Market Research

A prediction market is a speculative market that harnesses the power of the “wisdom of the crowds” for the purpose of making predictions. This concept was addressed in James Surowiecki’s 2004 book how a diverse group of individuals is able to make decisions and predictions better than individuals in isolation or even than experts. The advantage of this method is that it doesn’t rely on asking individuals to make predictions about what they will do in the future but rather what they think other people will do.

Why do this? Research has shown that consumers are unreliable predictors of their own behavior. Consider, for example –

  • How New Coke failed at retail in the 1980s despite positive concept test results. (Source: Malcolm Gladwell’s bestseller Blink: The Power of Thinking Without Thinking, 2005).
  • That 71% of Americans say they would purchase a product that supports a cause over one that does not. Yet, why is it that the red iPod, which contributed to the Global AIDS Fund, was not the top-selling product from Apple®? (Source: Fast Company, Feb. 6, 2014).
  • In a Pew Research Center study in 2006, 5% of American respondents said that they themselves were obese, but they estimated that 37% of other Americans were, much closer to the 31% that was the reality at the time of the survey.

As social animals, consumers are good at noticing what other people are doing, sensing why they might be doing it and predicting what they will do. Therefore, a crowd can successfully predict the future if they have contextual understanding of the category where the prediction needs to be made.

In our study, 1,000 respondents (also known as traders) were recruited from a general population sample (i.e., crowd) who self-selected into an online survey based on a baseline understanding of personal finance. Traders were first given a detailed description of the new product. They were then introduced to a number of consumer profiles (e.g., vignettes) one at a time, representing possible audiences for the new product. In total, six audience profiles were developed based on consumer segmentation and extensive secondary research, leveraging agency resources (i.e., Mintel, MRI). The vignettes included a narrative, describing audiences as people, and photography was used to bring them to life.

Next, traders were asked to invest in the profiles that they thought were most likely to purchase the new product, and to explain the reasons behind their predictions. A prediction market is like an online stock investing game where traders “invest” virtual dollars in ideas, products, or assets to be tested. Based on the investments they make, traders have the potential to win a greater incentive than the incentive they earn through completing the survey alone … if they invest in the winning idea. This “skin in the game” approach is a critical component of the methodology as it fosters engagement and thoughtfulness on the part of traders.

The Outcome

Three audience segments were clearly identified; one that was completely unexpected for the client. One segment that did not rank high for the new product was the client’s core audience for its main product line. Media/engagement touchpoints and messaging insights were also determined.