Having Expertise Is Not the Same as Using It

On the evening before the Space Shuttle Challenger launch in January 1986, engineers at Morton Thiokol were worried about the unusually cold weather. Their concern centered on the O-rings that helped seal joints in the shuttle’s solid rocket boosters. Previous flight experience had given them reason to worry about how those seals would perform at lower temperatures, and during a teleconference with NASA the engineering team initially recommended against launching below 53 degrees Fahrenheit, the lowest temperature in the available flight experience.

The recommendation did not survive the decision process intact. After NASA representatives challenged the basis for the temperature limit, Thiokol managers caucused privately and reversed the company’s position. The Presidential Commission investigating the accident later found that the people who ultimately approved the launch were not aware of Thiokol’s initial written recommendation against launching below 53 degrees or of the engineers’ continuing opposition after management changed the recommendation.

The following morning, Challenger broke apart 73 seconds after liftoff, killing all seven crew members. The Commission identified both the technical failure of the O-ring seals and serious weaknesses in the launch decision process. For the purposes of organizational reasoning, one feature of the case is especially important: NASA and Thiokol did not simply lack engineering expertise. Relevant specialists existed, examined the problem, and raised a material concern. What changed was the form in which that judgment reached decision authority.

That distinction matters well beyond aerospace. Organizations invest heavily in expert people, but the value of that expertise depends on what happens after the expert notices something important. Knowledge can be filtered, compressed, reframed, delayed, separated from its uncertainty, or converted into an apparent consensus as it moves through the organization. The expert can remain employed, available, and technically correct while the decision is made without the full benefit of what the expert knew.

Expert Knowledge Can Be Lost in Transmission

Organizations often think about expertise as an asset they possess. They hire engineers, physicians, attorneys, operators, researchers, analysts, security specialists, and people with years of accumulated experience. Once those people are inside the organization, it is tempting to assume that the organization therefore has access to what they know.

The Challenger case shows why that assumption is too simple. A technical judgment can be communicated and still fail to reach the eventual decision in a form that preserves its meaning. A conditional conclusion can become unconditional. A dissenting judgment can disappear inside a management recommendation. Uncertainty can be removed from a summary. A specialist concern can be translated into schedule, cost, or program language that changes how seriously it is treated. None of those mechanisms requires the expert to be literally silenced; the loss can occur through ordinary organizational transmission.

This means a useful expertise system has to preserve more than the expert’s bottom-line answer. Decision makers may need to know how strongly the conclusion is supported, which conditions it depends on, where disagreement remains, what evidence is missing, and what would cause the specialist to change the recommendation. If those relationships disappear while the conclusion travels upward, the organization may preserve the words while losing the knowledge that made the words meaningful.

Columbia: A Different Expertise Failure

Seventeen years later, the loss of Space Shuttle Columbia exposed a related but different failure. During Columbia’s January 2003 launch, insulating foam broke away from the external tank and struck the orbiter’s left wing. Engineers reviewing the launch imagery began trying to determine whether the impact could have caused dangerous damage, but the available imagery was not good enough to resolve the question confidently.

The Columbia Accident Investigation Board documented multiple attempts by engineers and managers to obtain better on-orbit imagery. Members of the Debris Assessment Team believed additional images were important to reduce uncertainty about the strike. Yet the requests moved through informal and sometimes unconventional channels, were treated in part as an engineering desire rather than a critical operational need, and were ultimately cancelled. The Board also found that Shuttle managers were not actively directing the Debris Assessment Team, did not regularly engage its leaders about assumptions, uncertainties, interim results, or the value of additional imagery, and did not bring more experienced analysts into some of the work.

The expertise problem was therefore not identical to Challenger. In Challenger, a relatively explicit engineering judgment was altered before it reached final decision authority. In Columbia, specialist analysis, uncertainty, evidence needs, and concern were never integrated adequately into the decision environment in the first place. Engineers continued to investigate possible failure scenarios, but some of those concerns never reached the Mission Management Team that held operational authority over the mission.

The contrast is useful because it shows that ‘listen to experts’ is not a sufficient organizational prescription. The expert has to become connected to the right problem, receive enough context and evidence to apply expertise well, communicate uncertainty and disagreement faithfully, and participate in a process that can integrate specialist knowledge with the larger decision before the commitment becomes irreversible.

Collective Knowledge Is Not the Same as Collective Use

Research on group decision making shows that this problem is not limited to large hierarchies or catastrophic events. Hidden-profile studies deliberately distribute relevant information across group members so that no one person initially has everything needed to identify the best choice. The group can reach the strongest answer only if members surface and combine the information that is uniquely held.

A meta-analysis by Li Lu, Y. Connie Yuan, and Poppy McLeod summarized 65 hidden-profile studies involving 3,189 groups. Across those studies, groups mentioned substantially more information that members already shared than information uniquely held by individual members, and groups working with distributed information were about eight times less likely to identify the correct solution than groups in which all relevant information was already shared. Better pooling of unique information was associated with better decision quality.

The implication is narrower than saying that meetings are ineffective. It is that collective possession of knowledge does not automatically become collective use of knowledge. A team can contain enough information to make a sound decision while ordinary discussion reinforces what participants already have in common. Adding another knowledgeable person to the room does not solve the problem if the process still fails to bring that person’s distinctive evidence into the reasoning.

Research on transactive memory systems addresses another part of the same issue. Teams do not need every member to know everything. They can distribute specialization while developing a shared understanding of who knows what, whether that knowledge is credible, and how the team coordinates access to it. Kyle Lewis’s field research developed and validated measures of transactive memory across laboratory teams, MBA consulting teams, and technology-company teams, treating specialization, credibility, and coordination as parts of a team-level knowledge system.

That perspective shifts the organizational question. It is not enough to ask whether someone with the right title or credential is present. The organization needs a workable way to recognize where relevant expertise resides, connect it to the problem, and allow different forms of specialized knowledge to be combined without requiring every participant to become an expert in every domain.

Apollo 13: Designing for Integration

NASA also provides a contrasting example of how differentiated expertise can be organized for use. When an oxygen tank ruptured aboard Apollo 13 in April 1970, the planned lunar landing became an effort to keep the crew alive and return them safely to Earth. The problem crossed propulsion, electrical power, navigation, life support, communications, consumables, procedures, spacecraft configuration, and crew operations. No single individual could have been the deepest specialist in all of those areas.

Mission Control was designed around that reality. Flight controllers in the main control room were responsible for particular systems and operational functions, while larger groups of specialists in supporting rooms worked detailed technical problems and fed their analysis into the front-room decision process. Flight Directors retained responsibility for integrating the mission state and making operational decisions rather than attempting to replace the specialists surrounding them.

Apollo 13 should not be reduced to a claim that organizational structure alone produced the successful return. The outcome depended on technical preparation, improvisation, spacecraft capabilities, crew performance, leadership, and extensive work across NASA and its contractors. The case is useful for a narrower reason: it demonstrates an architecture in which specialized knowledge can be developed in parallel and routed toward an integrating decision function without requiring the decision maker to possess every specialist method personally.

That is the opposite of treating expertise as something that matters only when the senior decision maker happens to know the right answer. A strong system makes expertise reachable, preserves specialist boundaries, and creates relationships through which different forms of knowledge can influence the whole problem.

Designing the Path From Expertise to Decision

Taken together, these cases suggest that expertise has an organizational path. It has to exist, but existence is only the beginning. The organization has to recognize where relevant expertise may sit; the right people have to become connected to the problem; they need enough context to understand what their expertise implies; their analysis has to retain uncertainty, conditions, and dissent as it moves; and different specialist perspectives have to be integrated without pretending that one specialty owns the whole decision.

Authority and timing are part of the same path. Specialists need enough standing for material concerns to remain visible when they challenge the prevailing interpretation, but specialist expertise does not automatically confer authority over the broader organizational commitment. The people who legitimately own the decision need to receive the relevant expert reasoning in a form they can evaluate, and they need to receive it while there is still time for the information to change what happens.

Failure at any point can make expertise effectively unavailable. The organization may never identify the right person. An expert may be consulted too narrowly. Their analysis may be stripped of uncertainty. A manager may convert disagreement into apparent consensus. Several specialists may answer their individual questions correctly while nobody owns the relationships among those answers. Or the knowledge may arrive only after a launch, contract, investment, or other commitment has become difficult to reverse.

Those are not simply hiring problems. They are failures in the architecture through which expert knowledge becomes organizational judgment. This is why the statement ‘we had an expert involved’ tells us very little by itself about whether expertise actually influenced a decision.

The Real Asset Is the Path

Organizations should continue investing in expertise. The harder question is whether they invest with equal care in the relationships that allow expertise to matter. Challenger shows that a specialist judgment can be expressed and still reach decision makers in a materially altered form. Columbia shows that expert analysis and uncertainty can remain insufficiently connected to the people responsible for the commitment. Hidden-profile research shows that groups can collectively possess information they fail to use, while transactive-memory research helps explain why recognizing and coordinating specialization is itself a capability. Apollo 13 illustrates a system deliberately organized to integrate differentiated expertise around a whole-mission decision.

The practical test is therefore not only whether the organization has knowledgeable people. It is whether the decision system can preserve and use what those people know. When a consequential decision goes wrong, one of the most revealing questions may be what happened to the relevant expert knowledge between the moment it became available and the moment the organization committed to action.

Selected sources and further reading

Rogers Commission (1986), Challenger launch decision and O-ring findings | Columbia Accident Investigation Board (2003), Volume I | Lu, Yuan, and McLeod (2012), hidden-profile meta-analysis | Lewis (2003), transactive memory systems | NASA, Apollo 13 history