News Stay informed about the latest enterprise technology news and product updates.

Software taps into the zeitgeist to predict the future

Move over, wikis. Companies are using prediction markets to get a read on what employees think about business outcomes and how often they're right.

Inkling Inc., a Chicago software company, is betting that the mentality of the masses trumps the wisdom of a few. Inkling is one of a growing number of vendors that sells prediction market software to corporations. The premise behind the software is that if you give a large number of people a risk-free betting exchange, they will generate a more accurate prediction of the future than a small number of specialists.

In the corporate equivalent of a trend sweeping everything from the presidential primary elections to box office receipts, companies like Cisco Systems Inc., The Procter & Gamble Co. and General Mills Inc. are tapping the anonymous viewpoints of their employees to quickly get a handle on the business zeitgeist. Employees are asked to bet on issues ranging from whether a store will open on time to how much the public will pay for a particular product to the likelihood of a competitor having a banner year.

"Essentially, these prediction markets help expose information that would not otherwise make it itself available," said Adam Siegel, founder and CEO of Inkling.

Large companies tend to be hierarchal and highly politicized, said Siegel, who worked at consulting firm Accenture Ltd. before founding Inkling in 2006. The political culture has "all sorts of ramifications" for decision making, forecasting and risk mitigation.

"We would see this every time we went in on a consulting job. The lower levels don't understand how decisions are being made at the upper levels. The upper levels didn't know what the lower levels were doing. There were pockets of information all over the place that were not being utilized," Siegel said.

Inkling is being used to tackle a wide range of problems relating to defining strategy, reducing risk and improving forecasts of key performance metrics. The prediction markets are used as an early indicator of looming risks, as input for strategic decisions and as a filter to test the viability of new ideas.

Inkling offers the following example on its website of how its platform works. If the aim, let's say, is to predict how many articles Wikipedia will have in English at the end of the year, a corporation would begin by setting a value then asking employees to bet on that value: Is it too high, too low or just right? The users' bets cause the original prediction to come more in line with their beliefs. The company might then make use of this information to alter its strategies and follow the trends in the market to see if its course corrections are having the desired effect.

Unlike when companies do surveys, where a questionnaire is rolled out and analyzed, and another survey done perhaps in six months, predictions markets are meant to be an ongoing marketplace of questions. If a company asks users what they believe the quarterly sales number for the quarter will be, it would keep the market open until the end of the quarter, when the result is known. The company would then liquidate the market to reward people who were right and perhaps punish people who were not so right, but typically also have other questions going on at the same time.

"The benefit to the company is they have this ongoing marketplace where people are giving their opinion about what they think is going to happen," Siegel said.

San Jose, Calif.-based Cisco, for example uses the Inkling platform for quality assurance. Every time Cisco builds a new iteration of its routers, there are potential delays and snafus related to hardware and software components. Such interruptions cost millions of dollars. Cisco production teams are running markets on the potential flaws.

"They went back and looked at all the things that bit them in the past and could bite them in the future and are running markets on them to ask what the probability of those things happening is," Siegel said.

Cisco is also using Inkling for competitive intelligence on the assumption that its tens of thousands of employees who work in Silicon Valley have their ears to the ground on what their competitors are doing or how they are doing.

Market manipulation

But couldn't traders manipulate the market -- for example, employees bet against a project coming in on time, so the powers that be push back the date? Siegel said the challenge for companies is getting a diverse group of people to participate.

"If you are asking a question about sales, you never ask only the salespeople themselves because typically they have a similar opinion. So you involve marketing people and product development executives," Siegel said. "When you have a diverse audience it is hard to get all of them to collude, because those groups are naturally competitive internally, most of the time. The challenge that you have is getting people who don't all have a dog in the fight."

The market rewards for being right, he said, not for being part of the crowd. If you care about making the most money, you will care most about making the right prediction, not by going along with other people. He gives the example of the traders who bet against the housing market when it was still booming and are now making millions of dollars on its collapse.

"It doesn't always pay to go along with the herd mentality. Also, all the questions asked are quantitative, not qualitative," he said. In other words, the questions that corporations test reference outcomes that will eventually happen. "There is a day of reckoning at some point, where, for example, either that product did or did not well, or met the cost-benefit analysis or not. It will be aright and wrong at the end."


Siegel said Inkling has worked hard to make its trading interface easy by utilizing something called an "automated market maker." Here is how the company website explains it: In a real stock market, when you want to sell shares, there has to be a buyer. Likewise, when you want to buy, someone needs to be selling. The price of the stock is set by the supply and demand of the shares being bought and sold. If more people want to buy, the stock price goes up. If more people are selling, the stock price goes down. So, for example, if the user market drives down the share value of a new product from the starting share price of $80 to $20, the company might want to rethink that new product.

By using an automated market maker, Inkling claims to remove the complexity of this "auction" process by always being the buyer and seller of shares. Inkling determines demand by the amount being bought and sold and sets the price accordingly. This also allows you to have far fewer people participating in a market and still have it yield statistically relevant results, the company states, citing research by Jed Christiansen, a graduate of the London School of Economics, who used Inkling Markets as his research platform and found that as few as 16 people participating in a market consistently delivered accurate predictions.

Inkling's algorithms are driven by the work of Robin Hanson of George Mason University, a pioneer in the field of prediction market.

Group think, by the drink

Inkling offers a Web 2.0 approach, allowing companies to subscribe to the software over the Web. Users can set up public decisions markets or create their own private markets.

It doesn't always pay to go along with the herd mentality.

Adam Siegel, founder and CEO. Inkling Inc.

Gartner Inc. analyst Mark Raskino recently singled out Inkling as a "cool" vendor, stating that the company "is making the relatively new but powerful management tool of 'prediction markets' easily accessible to a wider audience. They have made corporate and personal entry-level experimentation trivially simple and low cost, allowing people to get started quickly," Raskino said.

Prediction markets are not new and their very first corporate use can be traced back to the early 1990s, Raskino points out in a recent report. "However, it is only since popularization by James Surowiecki in the 2004 book The Wisdom of Crowds that this idea has really started to capture the imagination of mainstream management thinkers," he said.

Though the prediction market is becoming a more popular and well-known business idea, Raskino cautions it is still in the early-adopter phase. Other new companies in this field include Consensus Point LLC, NewsFutures Inc., Intrade and Qmarkets.

Let us know what you think about the story; email: Linda Tucci, Senior News Writer

Dig Deeper on Digital business management

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.