No one ever got fired for buying IBM

“No one ever got fired for buying IBM” is a workplace slogan that crystallizes the appeal of the safe, conventional choice: when a decision fails, following the crowd can seem less blameworthy than taking an independent risk. The book uses it to illuminate herding across professional life, from football coaching and investment to organizational decision-making, and to show why institutions need ways to preserve independent information while bringing it together.

The safety of the crowd

The slogan appears in a discussion of risk aversion among National Football League coaches. Going for it on fourth-and-two may make strategic sense, but it exposes a coach to a public, unmistakable failure; punting limits the immediate possibility of disaster. When peers all use the same strategy, adopting a different one can carry a professional cost even if the alternative is better in the long run. The result is a choice to fail conventionally rather than risk failing conspicuously: the kind of herding captured by the reassurance that no one gets fired for buying IBM.

The same pressure applies beyond football. Money managers are employed on the promise that they can outperform the market, yet many follow competitors into the same investment strategies and stocks. Their imitation makes an already difficult task harder, but it can still look safer than standing apart and being wrong alone. The slogan therefore names more than a preference for a familiar product: it expresses how social proof and fear of individual blame can override the incentive to think independently.

Independent judgment and information cascades

The book’s classroom experiments show why choices made in sequence can make conformity self-reinforcing. In an experiment by Angela Hung and Charles Plott, participants privately drew marbles from one of two urns: one contained twice as many light as dark marbles, the other twice as many dark as light. Each person guessed which urn was in use, while later participants could observe earlier guesses but not the colors they had drawn. Faced with a private signal that conflicted with the visible sequence of guesses, students usually followed the group; information cascades began in 78 percent of the trials.

Hung and Plott then changed the incentive. Instead of paying each student for an individually correct answer, they made payment depend on whether the group’s majority answer was correct. Students consequently had reason to attend more closely to their own private information, even though that information was imperfect and could lead an individual to guess incorrectly. The group’s collective judgment became significantly more accurate than that of the cascading groups. The contrast gives the IBM slogan a practical counterpoint: independent judgments may look less safe one by one, yet protecting them can make the group smarter. More generally, the book argues that organizations should, where possible, elicit judgments simultaneously rather than let early choices steer those who follow.

Coordination without suppressing local knowledge

The tension is not simply between central direction and individual autonomy. Decentralization allows people close to a problem to use tacit knowledge grounded in a particular job, place, or experience. It encourages specialization and independence, but information discovered in one part of a system may never reach the rest. The challenge is to make local knowledge collectively useful without erasing its local character; aggregation, rather than centralized command alone, is presented as the answer.

The intelligence community illustrates the stakes. After the September 11 attacks, Congress emphasized failures of information sharing: agencies had accumulated localized snapshots without forming a coherent picture of the threats. Senator Richard Shelby criticized the FBI’s decentralized structure for leaving information in largely independent fiefdoms, and argued for an agency above the disputatious bureaucracies. The book sets that proposed centralization against a different diagnosis: the community needed better ways to aggregate what its separate parts knew, not simply to replace their independence with top-down control. Two years after the attacks, it still lacked a unified watch list drawing on data from all agencies; linking databases could have improved information flow while preserving agency autonomy.

Incentives, cooperation, and organizational design

The slogan also belongs to a wider account of how incentives shape whether people share information or contribute to collective goods. In public-goods experiments by Ernst Fehr and Simon Gächter, each of four players begins with twenty tokens and chooses over four rounds whether to contribute to a shared pot. A contribution costs its maker one token but earns 0.4 tokens for each member of the group, creating a collective return that exceeds the individual cost. Most participants initially contribute about half their tokens, but contributions fall as players observe free riding; by the final round, 70 to 80 percent are free riding. As with herding, willingness to cooperate depends partly on what people see others doing.

Organizations can also reward the appearance of safe performance at the expense of honest information. Donald Roy, working as a lathe operator in a machine shop in the early 1950s, observed a piece-rate system with successive production thresholds and a cap. Workers had reason to restrain their output: exceeding a target could lead the company to raise the hurdle, so they learned to manipulate the rate rather than simply produce as much as possible. The book connects this problem to corporate targets and bonuses, which can encourage managers to lowball forecasts and use accounting maneuvers to meet expectations. Together, these examples show why a slogan about choosing the option that protects one’s job resonates: institutions can make conformity and strategic misrepresentation individually safer, even when they undermine the information and cooperation that better collective decisions require.