Risk think tank: Prediction market platform taps the wisdom of employees, partners, customers

Anticipating market response can be a virtual guessing game when it comes to launching a new product or determining key features during product development. Traditional surveys offer some insight to customer opinion, but usually their accuracy is questionable since the response rate is typically low.
By Karen Dilger, contributing editor (kadilger@comcast.net) March 1, 2008

Anticipating market response can be a virtual guessing game when it comes to launching a new product or determining key features during product development. Traditional surveys offer some insight to customer opinion, but usually their accuracy is questionable since the response rate is typically low.

That’s why prediction market solutions, which allow a group of people to express their views over time, represent the latest innovation in market research for many companies. Such solutions can improve forecasting, expose quality problems early, and uncover risk in the supply chain.

“Knowing what could happen always makes an organization better prepared,” says Adam Siegel, CEO and cofounder of prediction market specialist Inkling Markets . “And it’s common sense to take advantage of the collective wisdom that’s already present in your company.”

Using such a solution, it’s possible to predict answers to questions by buying and selling “shares in stocks” representing the possible answers. It works first by asking a question like, “Which book do you think will sell the most copies?” The books would be represented as stocks that people buy shares in.

If “Book 1” has a stock price of $40, that means “the crowd” thinks there is a 40-percent chance that book will sell the most. When people buy shares in that book, the price goes up; and when they sell shares in that book, the price goes down.

“The market does not represent just a snapshot of time, but conditions may fluctuate and change over time as players continue to trade,” says Siegel. The perpetual marketplace, which functions as Software-as-a-Service, runs continuously for a certain time period and becomes part of the daily routine.

Adam Siegel, CEO and cofounder of prediction market solutions supplier Inkling Markets, says the creativity and anonymous nature of a prediction market attracts respondents that would normally shy away from or ignore a traditional survey.

Proponents say the creativity and anonymous nature of a prediction market attracts respondents that would normally shy away from or ignore a traditional survey. “Respondents are more likely to say exactly what they think if they answer questions anonymously,” says Siegel. “Plus, having fantasy currency—and buying and selling shares—makes it a fun, interactive experience.”

Siegel cites a real-world example of a router that contained a design glitch. Employees, including engineers, were asked about the number of critical bugs they thought existed in the product. “Some knew where the risks and weaknesses were in the router, and the company was able to narrow down the problem areas,” says Siegel.

“An engineer might be hesitant to point out a problem when he or she feels the pressure of a looming product release date,” he adds. “However, the prediction market creates an easy outlet to report their thoughts without giving their name.”

Hoping to improve its forecast accuracy, Chrysler is using Inkling Markets in its volume-planning department in Auburn Hills, Mich., during release of its new minivans. Although the company has its own internal estimates and market forecasts, it wanted to see how well the official forecast compared with prediction market results.

Running the market for four weeks, the questions focused on the percentage mix of the three available seat options: standard bench seats, stow and go, or swivel and go.

Prediction markets software tools help large organizations improve forecasting, accelerate innovation, and better allocate resources.

“Minivans are purpose-driven vehicles,” says Brian Wallace, manager of strategic planning for Chrysler. “Buyers are more rational than passionate. If we don’t have enough of the upgrade version, such as the swivel and go, then we would have to sell the standard version for $500 less.”

Chrysler invited 235 people to participate in the market, and received a 25-percent response rate across all levels of seniority.

“Dealer orders gave us a more accurate forecast than the prediction market, but we were not surprised by that since we receive the dealer report a month early,” says Wallace. The company is considering doing another prediction market to test the results with another product line.

Although the jury is still out on the accuracy of prediction markets, they still remain a very popular research tool, says Paul Metz, senior VP for CNR Research Services , Chicago. “At a minimum, prediction markets are different, and today’s businesses are clamoring for innovation and efficiency.”

Metz says accuracy of prediction markets seems to be in line with scientific surveys. “Traditional surveys are flawed because national surveys or polls are always off the mark,” says Metz. “So the question remains, if a prediction market has a different answer than a traditional survey, which one is flawed?”