InnoCentive

InnoCentive, a spinoff of Eli Lilly, is presented as the setting for an experiment in collective forecasting: could a diverse group of employees, trading on information about prospective medicines, distinguish drug candidates likely to win U.S. Food and Drug Administration approval from those likely to be rejected? The market quickly separated the six candidates in the expected direction, making InnoCentive a concrete example of how decision markets can bring dispersed judgment to a consequential corporate problem without making the market itself the final authority.

The corporate problem

Drug development presents companies with a high-stakes forecasting challenge. A pharmaceutical company’s profits depend on developing successful drugs while limiting investment in unsuccessful ones, so a reliable way to identify likely approvals in advance could be valuable. The example places InnoCentive’s experiment within a broader question about how companies make difficult decisions when the relevant information is spread across people and conventional organizational processes can obstruct its use.

The book distinguishes gathering information from combining judgments about what that information means. In its account of the U.S. intelligence community, agencies possessed specialized knowledge but lacked a mechanism to aggregate judgments across them; simply putting a central authority in charge could improve information sharing while sacrificing the diversity and local knowledge that specialization provides. The corporate counterpart is not to abolish expertise or formal authority, but to create ways for independent perspectives to contribute to a common forecast.

The drug-candidate market

InnoCentive designed realistic profiles and experimental data for six different drugs. The outcomes were already known to the experiment’s designers: three candidates would be approved and three would be rejected. That setup made it possible to test whether participants could distinguish the two groups through market activity rather than merely offer opinions in a meeting.

A diverse mix of employees made up the market. When trading opened, prices for the candidates that would be approved rose quickly, while prices for those that would be rejected fell. The reported result is a clear directional separation between the two groups: the market identified the eventual winners and marked down the eventual losers. The account does not provide individual prices, participant numbers, or a separate accuracy percentage for this experiment, so its evidence is the rapid contrast in market valuations rather than a quantified comparison with another forecasting method.

What the example demonstrates

The experiment illustrates how a decision market can turn dispersed judgments into a legible collective forecast. In the book’s analysis, markets can help circumvent political infighting, sycophancy, and the confusion of organizational status with knowledge. Their anonymity and their production of a relatively clear result can give participants incentives to uncover and act on useful information, rather than orienting their conclusions chiefly toward pleasing a supervisor or defending an agency’s position.

InnoCentive’s example is also a limited experiment, not proof that markets settle every corporate question or guarantee successful strategy. Its six candidates had constructed profiles and data, and their approval outcomes were known to the people who designed the test. The result shows that a diverse employee market could sort these candidates in line with those known outcomes; the account does not establish that an equivalent market would resolve every uncertainty in live drug development.

Markets as advice, not authority

The argument drawn from InnoCentive is that major corporate decisions should be informed by decision markets, not made by them. Market prices can contribute an aggregated forecast, but they do not remove the need for accountable decision-makers. Indeed, the more important the choice, the less sensible it is to concentrate judgment in one person alone. Boards are formally meant to have the final say on major corporate decisions, although in practice they may defer to the chief executive.

Alph Bingham, InnoCentive’s chairman, expressed the contrast through a comparison with financial markets: people would find it strange if the chief executives of Goldman Sachs, Morgan Stanley, and Merrill Lynch decided what the prices of all companies’ stocks should be each morning, because a market is assumed to determine value better than a few people. Yet drug-company chief executives routinely decide which drugs to continue developing and which to abandon. The analogy does not say that market forecasts should exercise executive authority; it highlights the inconsistency between trusting collective valuation in one domain and reserving similarly consequential judgments entirely to individual executives in another.

Place in the book’s argument

InnoCentive supplies an organizational case for collective judgment within a larger account of how groups can use knowledge dispersed among their members. The example complements corporate forecasting at Hewlett-Packard, where a small internal market for printer sales outperformed the company’s forecasts 75 percent of the time over three years. InnoCentive shifts the application from sales forecasting to the selection of prospective drugs and shows market prices quickly separating candidates with different known outcomes.

The case therefore supports a specific institutional lesson: aggregate independent judgments while retaining formal human responsibility for the decision. It joins the book’s criticism of the assumption that the smartest individual must be the best decision-maker, but does not imply that expertise, leadership, or hierarchy should simply disappear. InnoCentive’s role is to show how a market can make distributed assessments visible and useful before executives decide what to do.