The Management Hat

On the evening of January 27, 1986, a group of engineers at Morton Thiokol in Utah joined a teleconference with NASA and argued against launching the space shuttle the next morning. The forecast for Cape Canaveral was freezing. The rubber O-rings that sealed the joints of the solid rocket boosters had never flown in weather that cold. The coldest previous launch had been 53 degrees Fahrenheit, and even then the seals had shown damage. The engineers' recommendation was plain: do not launch below 53 degrees.

What happened next is recorded in the Rogers Commission report, and it is worth reading slowly. NASA officials pushed back hard. One asked, in exasperation, when Thiokol wanted him to launch — next April? The company asked for an offline caucus. Inside it, a senior Thiokol manager turned to the head of engineering and told him to "take off your engineering hat and put on your management hat." The recommendation was reversed. The engineers who had argued the case were not asked to sign the new one.

Challenger lifted off at about 36 degrees. Seventy-three seconds later, it broke apart. Seven people died.

It is tempting to read this as a story about villains. It is more useful, and more uncomfortable, to read it as a story about a room. Nobody in it wanted a disaster. The schedule was public, the program was under pressure, and the burden of proof had quietly flipped. The engineers were no longer asked to show that it was safe to fly. They were asked to prove that it was unsafe, with data that by definition did not yet exist, because nobody had ever flown that cold.

That inversion is the most common way good organizations make bad decisions. It rarely announces itself. It sounds like realism: we need more than a feeling; show me the numbers; everyone else is comfortable. David Hume warned that the mind slides from what has happened to what will happen, as though the past were a promise. The shuttle had flown before with damaged seals, so damaged seals came to look like part of the design.

Weeks later, the physicist Richard Feynman dropped a piece of O-ring material into a glass of ice water during a televised hearing and showed that it lost its resilience. It took one clamp and one glass. In his appendix to the report he wrote the sentence that every board should keep in a drawer: "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled."

The lesson for anyone who chairs a meeting is not "listen to engineers." It is narrower and harder. Notice when the burden of proof moves. Notice when a dissent is being asked to meet a standard that the consensus never had to meet. And notice who has been asked to change hats.

The minutes of a decision are written by the people who stay in the room. Make sure the person who said no is still in it.