Watch session videos and thought leader interviews from the Advertising Research Foundation Re:Think 2010 conference.
Register now for the Audience Measurement conference on June 22-23 in NYC.
To paraphrase an old joke, what do you get when you put 10 market researchers in a room together? Eleven different opinions.
Those of us in the research world can argue all day about the best new research methodologies or the validity of NPS, but there is one thing we all can agree on: our industry needs innovation.
Research clients have asking us for innovation for years, but unfortunately the best we can give them is often just a new statistical analysis trick. They want more than that. They want fresh new approaches to research that help them make better business decisions, preferably more quickly and cost-effectively than today. I believe that we’ve stumbled upon just such an innovation: the prediction market.
Prediction markets are powered by “the wisdom of crowds,” a concept made popular by best-selling author James Surowiecki’s book by the same name. The premise of the book is that large groups of people, even non-expert people, are very often smarter than an elite few no matter how brilliant those few may be. The author provides numerous examples of when groups have proven better than individuals at solving problems, fostering innovation, coming to wise decisions and most notably predicting future outcomes.
One powerful mechanism for aggregating this “wisdom of crowds” is a prediction market – a virtual stock market that allows traders to buy and sell shares in future outcomes. Savvy researchers can track the movement of these share prices over time in order to accurately predict outcomes of interest to their clients, such as which of their new product concepts will be most successful or which advertising concept will have the greatest impact on sales.
Prediction markets have been used in other domains for many years with excellent results. For example, the Iowa Electronic Market (IEM) has been found to consistently predict election outcomes more accurately than the Gallup pool, the Hollywood Stock Exchange (HSX) predicts movie box office receipts with 93% accuracy, and Google even setup an internal prediction market to predict outcomes of interest to their top management even more accurately than their own full-time internal forecasters.
Now the market research industry is just starting to harness their amazing power. As it turns out, these markets allow us not only to predict concept success with astonishing accuracy, but also more quickly and affordably than traditional concept testing surveys, and with greater benchmarking capabilities.
Prediction markets can be setup using either internal employees or external consumers as traders. We’ve found that using external consumers is preferable when the market is being used to answer research questions, as you can avoid the challenge of maintaining high adoption levels in a hectic corporate environment.
One of the keys to setting up an accurate prediction market is allowing traders to self-select for only those markets they think they can win. Since in a prediction market the traders can only make money by investing in a winning concept, they have a natural incentive to participate in only those markets dealing with topics they are familiar with.
There are many other things to consider when setting up a prediction market, like whether to use real or fake currency, whether to allow traders to trade once or multiple times in the same market, how to resolve the market (i.e., pick a winner), and the best trading interface to show traders. My company spent 18 months experimenting with these variables and ultimately came up with a prediction market formula that works great for market research.
The world’s leading brands, companies like General Mills, Nestle, and Wrigley, are already using prediction markets for new product concept testing. Maybe you should as well.
