Why the Military Debriefs Everything, and Most Organizations Debrief Nothing

The United States Army debriefs everything. A live-fire rotation in the Mojave Desert, a maintenance cycle, a change of command, a training accident. Somebody sits the participants down, walks the event from beginning to end, writes down what the organization learned, and sends it to an office at Fort Leavenworth whose only job is to make sure the next unit does not pay for the same lesson.

Most organizations I walk into do not do this at all. The conference ends, the survey closes, the incident is resolved, and three people stand in a hallway and agree on what went wrong. Nobody writes it down. Two of them are gone by the next cycle, the third remembers it differently, and the same gap opens at the next event.

The difference is not discipline. It is method. The Army has one and most organizations do not.

Where the After Action Review actually came from

The common story is that the Army invented the After Action Review at the National Training Center in the 1980s. That is not quite right, and the real history is more useful.

The method was developed in the mid-1970s by the Army Research Institute for the Behavioral and Social Sciences, inside two tactical engagement training programs: SCOPES, fielded by the Infantry School in 1973, and REALTRAIN, fielded by the Armor School in 1975.⁶ What the National Training Center at Fort Irwin did, beginning with the first battalion rotations in October 1981, was scale it and make it permanent.¹ Observer teams delivered an initial review within two hours of the end of each mission. The method was codified into doctrine in 1993 as Training Circular 25-20, A Leader’s Guide to After-Action Reviews.⁴

Two features of that doctrine matter more than anything else in it, and both are counterintuitive to executives.

The first: an After Action Review is not a critique. TC 25-20 states it plainly. “A critique only gives one viewpoint and frequently provides little opportunity for discussion of events by participants.”⁴ The review is a professional discussion in which the participants discover for themselves what happened, not a briefing in which a senior person announces what happened.

The second: it does not grade. “An AAR does not grade success or failure. There are always weaknesses to improve and strengths to sustain.”⁴ The Army Research Institute traces this design intent directly back to S. L. A. Marshall’s combat interviews and describes the objective as a non-punitive atmosphere, contrasting the open climate of a review with the defensive climate of a critique.⁶

That is the whole trick. A debrief that assigns fault teaches the organization to stop reporting. A debrief that establishes what happened, in sequence, on the record, teaches it to get better.

The four questions

The Combined Arms Center version of the leader’s guide frames the review as four questions, and I have never found a better sequence for a civilian organization:⁸

1. Review what was supposed to occur.
2. Establish what happened.
3. Determine what was right or wrong with what happened.
4. Determine how the task should be done differently next time.

Notice what question one requires. You cannot review what was supposed to occur if nobody wrote it down. Half the value of a review is extracted before the event ever happens, which is why I treat planning alignment as its own engagement rather than as a preamble to the debrief.

The evidence that this works

This is not a military tradition that happens to be comforting. It is one of the better-measured interventions in organizational research.

Tannenbaum and Cerasoli meta-analyzed 46 samples covering 2,136 people and found that debriefs improve effectiveness over a control group by approximately 25 percent, an effect size of d = .67. Their conclusion was that organizations can improve individual and team performance by approximately 20 to 25 percent by using properly conducted debriefs. The effect held for teams and for individuals, in simulated and in real settings, and in medical and non-medical samples alike.⁷

Keiser and Arthur replicated it larger. Across 61 studies covering 915 teams and 3,499 individuals, they found an overall effect of d = 0.79, which they note is larger than some of the largest training method effects in the literature.⁵ Two things consistently moderated the result: whether the review was aligned to the individual or to the team, and whether performance was reviewed against objective media rather than memory. They also found that in military settings a highly structured review outperforms a loosely structured one, while in healthcare the two perform comparably.

The healthcare outcome data is real as well. A systematic review and meta-analysis in Intensive Care Medicine screened 2,720 studies, included 27, and found that 20 of them supported debriefing, with measured improvements in learning, non-technical skills, and technical performance. In cardiac arrest, debriefing was associated with a mean improvement of 6.80 in chest compression fraction and an odds ratio of 1.46 for return of spontaneous circulation. The authors are careful, and so am I: the effect on long-term patient outcomes remains uncertain.²

Debrief the wins, not only the losses

The single most common mistake I see is an organization that only debriefs disasters. Ellis and Davidi tested exactly that. Soldiers running successive navigation exercises improved significantly more when they were debriefed on their failures and their successes than when they reviewed only what failed. The authors also found that before the intervention, participants held far richer mental models of their failures than of their successes, and that the gap closed as they reviewed both.³

Read that again, because it is the finding that costs organizations the most money. Your people already understand why things break. What they cannot explain is why things work, which means they cannot reproduce it on purpose.

What I take from all of this

An After Action Review is not a meeting. It is a mechanism for converting one hard day into something the organization owns after the people who lived it have moved on. Four things have to be true for it to work, and every one of them is supported above.

It has to be facilitated by someone with no stake in the answer. Nobody in the room should report to the facilitator, and nobody should have a position to protect in front of him. It has to be structured, because structure is what separates a review from a conversation. It has to be measured against something objective rather than against what people remember. And it has to end in writing, with dates and owners, or the organization has held a therapy session rather than a review.

That is how I run it. A short confidential intake, one facilitated session, an assessment against four published criteria, and a ninety-day improvement plan delivered within seventy-two hours, followed by a leadership briefing and a thirty-day call to confirm the plan is actually moving. The buyer knows before the session exactly what will be assessed and on what basis, because the criteria are published rather than revealed afterward.

The Army figured this out fifty years ago and has never stopped doing it. There is no reason a hospital, a city, a university, or a church cannot do the same thing next month.