Xpree is a new Silicon Valley company that lets large companies make more efficient forecasts by tapping into the group wisdom of their employees.
The more I think about it, the more it seems crazy this hasn’t been done before.
Xpree lets a company’s employees bet on whether a product will, say, meet its launch date. Xpree encourages participation by awarding employees with perks and recognition by peers. While very early, Xpree is getting traction, having already signed up Electronics Arts and travel company EyeForTravel as customers.
Large companies have to hit deadlines, to meet market expectations. If they don’t, they get punished by Wall Street. From Microsoft in software (remember how the Vista took for ever), to Yahoo with advertising platforms (Panama, well, zzzzzzz) to Boeing's Dreamliner (11 months late and counting), all companies incur damage if their products are delayed, or if they aren’t any good when they do arrive. There’s a very large market need for better predictions.
Xpree lets companies gauge such things by tapping the wisdom of the people who know the most about the product — the rank and file who see the bugs that are likely to slow a product's sales, or conversely, the product's shine that will make it a major success. Xpree provides an online dashboard for employees to bet virtual money and accumulate it over time by being successful (see screenshot below).
There's a slew of other companies working on predictions markets — for example Industry Standard, Predictify, Hubdub and others — but they're consumer focused, and few of them serve corporations. Those that do serve companies — including Consensus Point, Inkling Markets (with seed funding from YCombinator) and News Futures — are in their earliest stages, like Xpree.
Xpree chief executive Mat Fogarty has a good start. He hails from Electronics Arts, where he was director of financial planning and reported to the chief financial officer on forecasting matters. He was consistently surprised that major public companies didn’t have better ways to forecast. If Electronics Arts released a new game, a few executives and a product managers might be responsible for estimating when it would be ready for release and how well it would do. But each of them had their biases, and they were often wrong.
It wasn’t just Electronic Arts that had problems; it's all companies, from giant conglomerates down to the smallest start-ups forecasting their burn rate (witness KnockaTV). Fogarty said he read the book Wisdom of Crowds by James Surowiecki and was inspired to change the paradigm of how forecasts are made in corporations.
A year and a half later, he’s come up with an online product that already has Electronics Art and two other companies as paid clients. For EA, some 240 employees are using Xpree’s betting platform to trade on the quality level of games. Xpree lets them bet whether a game will get a high or low rating on Metacritic, for example, which is a popular site that aggregates reviews from around the Web on games and other products. That way, if employees think a game will rate low or high, EA can adjust the way it pitches retailers on the game, or adjust its salesforce commissions.
But how do we know rank-and-file know more than the executives? Well, Fogarty says there’s increasing evidence suggesting that’s the case. In the initial trial, the crowd of employees at EA was more accurate than EA’s executive experts in forecasting game quality (more accurate in 10 out of 12 games). The average error of the crowd was half the error of the experts.
More and more, the largest U.S. companies are coming to the same conclusion. Google has its own predictions market, and 1,500 employees are using it. Internally, they’ve found that the longer an employee has worked there, the more likely they are to be accurate when betting (more here). Intel is using a market to predict the sale of its chips and reports that it has witnessed a 20 percent reduction in error compared to previous methods. The Hollywood Stock Exchange predicted 92 percent of Oscar winners in the last three years. Microsoft has experimented with a system to predict ship dates.
Notably, Fogarty (left in image above) says he’s getting more interest than expected from companies who want to use Xpree in the earliest stages of a product’s life, i.e, to help forecast how much it will cost to develop products (as opposed to forecasting delivery dates).
Fogarty runs a team of six, including Alam Kasenally (right), a former engineering manager in Yahoo’s mobile division and Geoff Tso (middle above) formerly a senior engineer at Siebel. They’re running on $270,000 in seed funding from friends and family, and are raising a first round of capital next quarter.
