Print this article | Return to Article | Return to CFO.com
Internal markets can solve thorny allocation problems and predict the future.
Alix Stuart, CFO Magazine
November 1, 2005
"Man is an animal that makes bargains," observed the great 18th-century economist Adam Smith. Today, this instinct is given free rein in markets around the globe. Most Americans — and certainly most public-company CFOs — take it as an article of faith that markets eventually lead to the fairest prices and most efficient distribution of goods.
Why, then, don't more companies resort to markets when it comes to making decisions about corporate resources? Market-based systems that let employees express preference and need through price have proven accurate, efficient, and technologically simple. "In principle, almost any company could use some kind of internal market," says Thomas Malone, a professor at Massachusetts Institute of Technology's Sloan School of Business.
But so far, few companies have large-scale internal markets under construction, according to Malone and others who consult in this area. NewsFutures Inc., one of the primary technology providers for such markets, says its client base has grown to 12 companies this year, up from 1 in 2003, and could hit a maximum of 50 over the next year. "I wouldn't call it a bandwagon yet," NewsFutures CEO Emile Servan-Schreiber says wryly.
Using internal markets within corporations is hardly an untested concept. One of the classic examples is the market BP Amoco set up in the late 1990s to reduce the oil giant's greenhouse-gas emissions. Each business unit was granted the right to generate one ton of carbon-dioxide emissions and given access to an electronic trading system that allowed it to buy more capacity or sell excess to other units. By 2001 — nine years ahead of schedule — the company had hit its targeted reductions after more than 4.5 million tons had been traded.
Researchers who work on such markets say there are still many allocation applications to be explored, from the mundane to the sublime. Bernardo Huberman and his group at HP Labs, the central research organization for Hewlett-Packard, has designed systems to let people buy and sell rights to computing power and even conference rooms. Malone has recently done extensive research for Intel on the use of a market for manufacturing capacity (as yet unimplemented), in which plant managers and salespeople could gain information from each other in order to minimize production costs and maximize profit margin. In theoretical runs, the allocations have come within 10 percentage points of perfection.
Another common use of internal markets is for eliciting predictions from employees about which product designs, prices, and sales strategies will be most effective. Hewlett-Packard, for example, gave salespeople 20 shares each and told them to buy and sell futures contracts based on how many printers they expected customers to buy at given prices. The results showed their expectations to be as close, if not closer, to actual sales as HP's official predictions had been.
Likewise, at Eli Lilly and Co.'s research labs, about 50 employees involved in drug development used an internal market to buy and sell shares of six mock drug candidates. In the end, traders had bid up the highest prices for the three drugs that ended up being the most popular on the market. Web-search firm Google, too, recently set up a system to let employees bid on likely launch dates for new products, with similar results.
The market method is "at least as accurate as, if not slightly better than, other methods" of gaining insight into products and customers, says Ely Dahan, an assistant professor at the University of California in Los Angeles. Besides the results he has found in his own research, Dahan points to the Iowa Electronic Market, which allows people to bet on the shares of votes each Presidential candidate will receive, and the Hollywood Stock Exchange, which allows (mock) bets on movie successes, as proof of the method. Both markets are typically more accurate than mere opinion polls, he says, in part because they let participants in on what others are thinking.
Another boon of internal markets: people may not need to be paid extra to play. NewsFutures's Servan-Schreiber and three co-researchers compared the real-money bets people made about 2003 National Football League outcomes through TradeSports.com with hypothetical bets about the same season made through NewsFutures.com. Surprisingly, they found that both markets were about as accurate when compared with actual results, and were much more accurate than individual predictions. "In the case of play-money," the researchers concluded, "knowledgeable players can be motivated, for example, by community bragging rights or by prizes awarded to the best forecasters."
The Control Thing
So why haven't internal markets become more common? Cost may be one reason, although likely not the primary one. NewsFutures charges from $25,000 for a short pilot project to $200,000 for a longer, more people-intensive one, since each project still requires some design and customization at this point. Of course, such tools also take training. Participants may need to be educated about what their bets mean or how to bid effectively for resources, and likely cajoled into using a new system.
A bigger obstacle, though, is the control thing. Most managers "assume that if there's a problem, the solution is to put some person in charge of it," says MIT's Malone. "If there's a problem, there should be someone to blame." But internal markets require that managers give up some control and let the market decide what to do, which is "scary" for some people even though "in many cases the outcome may be better and more in line with what managers actually want," he says. Malone, however, thinks the fear of losing control is largely overrated. "In reality, managers can still be in control of the process," he points out, "because they set the rules, and can intervene if necessary."
Another difficulty with internal markets is that employees may be reluctant to publicize what they know. "You might ask a question like, 'When can we ship a product?' through a market," says Dahan. "If the consensus of the managers is that you're going to be six months late, though, they're not going to want it known." In efforts to protect certain types of information, then, players may collude with one another in order to manipulate prices to show a different expectation. One potential solution to the collusion risk, Malone says, is to reward participants relative to their peers, rather than to their absolute profits in the internal market.
When it comes to resource allocation, HP's Huberman and his team have recently come up with a "truth-telling" mechanism within a reservation system for computing resources that helps others determine how likely those reservations are to be kept. Employees can choose a low-cost reservation with a high penalty for cancellation or pay a higher price upfront with minimal penalties for opting out later. "By choosing how much you'll pay and how much penalty you'll take, it reveals to us how likely you are to use your reservation," says Huberman, creating more incentive for employees to take only what they need.
No doubt there are more glitches to be fixed. But researchers are hard at work perfecting the concept. Meanwhile, if CFOs have any faith in Adam Smith, it may be worth their while to start thinking about how markets can help their companies — both inside and out.
Alix Nyberg Stuart is senior writer at CFO.
Four companies that have used internal markets to make decisions.
|BP Amoco||Business-unit managers traded greenhouse-gas emissions credits to reduce overall corporate pollution.|
|HP||Had sales team buy and sell futures contracts on potential sales volumes at given prices.|
|Eli Lilly||Asked scientists to trade stocks on six mock drugs to gauge which would be most successful.|
|Lets employees bid play-money on when new products and offices are likely to launch.|