After Action Review: How to Run One That Changes Results

Abstract review path from outcome to action

An after action review (AAR) is a short, structured conversation held soon after a task, milestone or event. The people who did the work compare what was meant to happen with what did happen, then agree what to keep and what to change. It is not an inquest and it is not a report for someone else. Done well, it takes about an hour, costs nothing, and gives a team a repeatable way to improve.

This guide covers the four questions, a step-by-step process you can run next week, how to handle the awkward parts (blame, silence, vague actions), and how to connect an AAR to customer feedback and wider business capability. It is written for owners and managers of small and mid-sized UK businesses, including founder-led professional services firms.

Strategic-concierge

Turn Reviews Into Lasting Capability

Strategic Concierge helps businesses improve execution and embed the capability needed to turn learning into sustained progress.

Explore Strategic Concierge

TL;DR

An AAR compares intended and actual outcomes with the people who did the work, and ends with specific changes, not a discussion summary.
The four questions are: what was supposed to happen, what actually happened, what were the differences, and what do we do differently next time.
Hold it shortly after a milestone, not only at the end of a project, and keep it to roughly 30 to 60 minutes.
It only works without blame. Name an owner and a date for every action, or the learning is lost.
Treat it as a management routine, not a one-off event, and link it to customer feedback and root cause analysis.

What an after action review is
The four questions
How to run an AAR: step by step
Timing: soon after, and more than once
Keeping it blame-free
Common mistakes
Using an AAR with customer feedback
A worked example
Our perspective
Where we can help
Turning lessons into lasting change
FAQ
Sources

What an after action review is

The NHS England learning handbook describes an AAR as a method of evaluation used when the outcomes of an activity or event have been particularly successful or unsuccessful, with the aim of capturing learning to avoid failure and promote success. The method was developed by the US Army, and the same handbook stresses that rank is left at the door so everyone has an equal chance to contribute.

Two points in that definition matter for a business:

It covers successes as well as failures. A project that went unusually well deserves the same scrutiny as one that went badly, because you want to repeat what worked on purpose.
It is task-based. An AAR looks at a specific piece of work, such as a client onboarding, a tender, a product launch or a campaign, rather than at “how the business is doing”.

The UK government’s lessons management guidance (annexes, September 2024) places AARs within lessons management: identifying lessons, prioritising them, implementing them and embedding them in practice. That last step is where most organisations fall down, and we come back to it below.

How it differs from a post-mortem and a lessons learned workshop

A traditional post-mortem tends to happen once, at the end, and often drifts towards finding who was at fault. An AAR is forward-looking and involves everyone who took part. Better Evaluation describes it as a way of assessing organisational performance by bringing a team together to discuss a task, event or project in an open and honest way.

A lessons learned workshop is the larger cousin. It usually covers a whole project, runs for a half day or more, and may include stakeholders from outside the team. Lessons Learned describes a workshop as an opportunity for project team members and stakeholders to get together and discuss lessons learned. In practice:

	After action review	Lessons learned workshop
Scope	One task, event or milestone	A whole project or programme
Typical length	30 to 60 minutes	Half a day or longer
Timing	Soon after the activity	At project close or major phase end
Participants	The people who did the work	Team plus wider stakeholders
Output	A short list of actions	A broader lessons log and recommendations

Most small businesses get more value from frequent short AARs than from one large workshop that is easy to postpone.

The four questions

The core of the method is four questions, set out by Better Evaluation:

What was supposed to happen?
What actually happened?
What were the differences?
What can we learn, or do differently next time?

Searchers often ask for “the five AAR questions”. Some templates add a fifth, such as “what should we sustain?” or “who will act on this?”. Both are sensible additions, but the four above are the standard core.

Question 1: What was supposed to happen?

This is harder than it sounds. Teams often discover that people held different expectations of the same piece of work. One person thought the goal was a signed proposal, another thought it was a good meeting. Write the intended outcome down in one or two sentences before anyone talks about results. If you cannot agree on the intent, that is already your first finding.

Question 2: What actually happened?

Stick to facts: dates, numbers, what was said, what was delivered. Use data where you have it, such as pipeline figures, delivery dates, response times or customer comments. Let each person describe their part of the sequence before anyone interprets it.

Question 3: What were the differences?

Place the two accounts side by side and ask why gaps appeared. Look at both directions: where you fell short and where you beat the plan. Keep asking “why” until you reach something the team can change. For recurring or serious problems, a structured method helps; our guide to root cause analysis methods explains when the 5 Whys is enough and when it is not.

Question 4: What do we do differently next time?

This is the question that justifies the meeting. Convert each lesson into an action that has an owner, a date and a way of checking it happened. “Communicate better” is not an action. “Send the client a one-page scope summary within 24 hours of kick-off, owned by the account lead, from the next engagement” is.

How to run an AAR: step by step

The NHS England handbook puts the session at around 30 minutes to an hour. Here is a process that fits that window.

Before the session

Choose the activity. Pick a specific, recent piece of work with a clear start and end: a pitch, a delivery phase, a campaign, a customer onboarding.
Invite the right people. Everyone who played a role, regardless of seniority. If the person who made the key decision is absent, the review will go around in circles.
Pick a facilitator. Someone who was not central to the outcome, or who can stay neutral. The facilitator keeps time, asks the questions and protects the ground rules. They should not be the most senior person in the room if that would quieten others.
Gather the evidence. Bring the original plan or brief, the key figures, and any customer feedback. Ten minutes of preparation prevents arguments about memory.

During the session

Set the ground rules (2 minutes). The purpose is learning, not grading. Nothing said is used in appraisals. Everyone speaks.
State the intent (5 minutes). Agree and write down what was supposed to happen.
Establish what happened (10 minutes). Build a short timeline with facts first, then perspectives.
Explore the differences (15 minutes). Identify what helped and what hindered. Separate causes from symptoms.
Agree actions (10 minutes). Limit yourself to the few changes that will matter most. Three well-owned actions beat twelve vague ones.
Close (2 minutes). Confirm who records the output and when the actions will be checked.

After the session

Write up a one-page summary the same day: intent, outcome, key differences, actions with owners and dates. Share it with the participants and with anyone who will run similar work. Then add the actions to wherever your team already tracks work. If they live only in a document, they will not happen.

The Guide to the After Action Review from CEBMa frames the same logic: teams that continuously assess their performance learn from both successes and failures.

Timing: soon after, and more than once

Practitioners generally recommend holding an AAR shortly after a milestone while the team is still together and memories are fresh, rather than saving everything for the end of a project. A review held three months after delivery tends to produce a tidy, edited version of events.

There is one nuance worth knowing when your review depends on customer opinion. Research reported by EurekAlert found that asking customers for a review as soon as possible is not always the best strategy, and that reminders cause problems when they arrive faster than customers typically take to form a view. So hold your internal AAR quickly, but give customers a sensible interval before you treat their feedback as the final word on how a service landed.

Keeping it blame-free

A no-blame culture is not a courtesy. It is what makes the data usable. UKHSA’s guidance on constructive debriefing and lessons identified is written for outbreak management, but the principle carries over to any team: people share what really happened only when they trust that doing so will not be used against them.

In a small professional services firm the risk is sharper than in a large organisation. Everyone knows everyone, the founder is often involved in the work, and a sharp comment can sour a relationship for months. Some practical safeguards:

Talk about decisions and processes, not people. “The scope was not written down” rather than “Sam didn’t write the scope”.
The most senior person speaks last. Otherwise their view anchors the discussion.
Ask what the situation made difficult. This surfaces system problems such as unclear briefs, missing tools or competing priorities.
Keep reviews separate from performance conversations. If people suspect crossover, candour disappears.

An external, neutral facilitator can help when the team is small, the stakes are high or the founder is part of the story. Outside facilitation is not essential for routine AARs, but it is worth considering for a major loss, a failed launch or a complaint that touched several people.

Common mistakes

Treating it as a final post-mortem. One review at the end of a project cannot change that project. Run shorter ones at milestones so you can apply lessons while the work is live.

Skipping the intent. Without a shared definition of success, the discussion becomes a debate about opinions.

Collecting lessons and doing nothing. The government guidance is explicit that lessons need to be embedded, not just recorded. A lesson that is written down but not acted on is only a note.

Inviting too many people. Beyond about eight to ten, quieter voices drop out. Split the review by workstream if necessary.

Reviewing only failures. If you only review when things go wrong, people learn to associate the AAR with trouble, and you miss the chance to copy what works.

Letting actions float. Every action needs one named owner, not a team, and a date.

Using an AAR with customer feedback

Many businesses collect feedback and then file it. Research on how service providers manage the customer’s experience across the journey, from Naples Forum on Service, points to uncoordinated management of that experience as a problem: activity happens in pockets without being joined up. An AAR is a simple way to join it up.

Try this: make customer comments a formal input to question 2. Instead of keeping complaints and compliments in a separate marketing or support report, bring them into the review and ask where what the customer saw differed from what you intended. Typical outputs include a changed handover step, a clearer expectation set at the sales stage, or a revised follow-up cadence.

The same study cautions that opportunistic, insistent sales tactics are destructive to the customer’s experience. That is a useful discipline for any AAR on a sales process: if the review concludes “we should have pushed harder”, test that conclusion against what the customer actually experienced.

If your reviews concentrate on sales outcomes, a win loss analysis is a natural companion, and our customer success playbook shows how to turn what you learn into retention routines.

Using an AAR with customer feedback — overview diagram

A worked example

Imagine a six-person consultancy that has just delivered a discovery phase for a new client, and the client has asked for a longer timeline than agreed.

Intended: complete discovery in four weeks, with a signed-off findings document at the end.
Actual: discovery took six weeks, and the findings document needed two rounds of revisions.
Differences: the client’s data arrived in week two instead of week one; the scope did not say who at the client signs off; the team started analysis before confirming the questions to answer.
Actions: add a data-readiness checklist to the kick-off (owner: delivery lead, in place for the next engagement); name a sign-off contact in every scope document (owner: founder, from the next proposal); hold a 30-minute question-framing session before analysis begins (owner: project lead, next project).

None of those actions blames a person, all are small, and each prevents a repeat of the same delay. That is what a good AAR looks like.

Our perspective

We see the same pattern in the businesses we work with: the review itself is rarely the problem. The problem is what happens afterwards. A team holds a thoughtful AAR, writes down good lessons, and three weeks later the same issue appears in the next piece of work because nothing in the way the business operates has changed.

That is why we treat the AAR as one part of building capability, not as a stand-alone meeting. Our approach moves through Build, Iterate, Optimise, Embed and Grow. An AAR fits at the Iterate and Optimise stages, where you are testing what works and tightening execution, and it feeds directly into Embed, where an agreed change becomes a standard part of how the business runs: a checklist, a template, a changed handover, a clearer role.

Three things we would tell any owner starting out:

Start with one recurring activity. Choose something you do often, such as onboarding a new client. Frequent repetition gives you quick evidence about whether a change worked.
Fix sequence before speed. If you have more lessons than capacity, decide what to change first. Our piece on choosing what to build next covers sequencing commercial work.
Judge before you act. A lesson is a hypothesis about what to change. Think it through before building anything around it, which is the argument in judgement before execution.

An AAR is cheap, and that is exactly why it is often neglected. The value comes from the discipline of doing it routinely and acting on it.

— Chris

Where we can help

If your reviews keep producing good insight and little change, the gap is usually capability: no clear owner, no agreed way of working, no routine to hold the change in place. We help small and mid-sized businesses and founder-led professional services firms close that gap. Our Foundation stage gives you a launch-ready base, and our Managed Capability plan provides ongoing support, from £999 per month, £2,850 per quarter or £10,000 per year. You can see the full options on our pricing page, read how the staged method works, or book a discovery call to talk through where your business is stuck.

Turning lessons into lasting change

The final test of an AAR is whether the next run of the same activity is better. A few habits make that likely:

Keep a single log of lessons and actions. One shared page is enough. Review it at your regular management meeting.
Check actions after a fixed period. Thirty days is a practical default. Ask: was it done, and did it help?
Change the artefacts. Update the template, checklist, brief or process document so the lesson is built in. People forget lessons, but a checklist persists.
Share across teams. A finding about client onboarding is probably useful to whoever manages renewals.
Look for patterns. After five or six reviews, recurring themes show you where a deeper change is needed. At that point, look at the structure of the business, not only individual tasks. Our guides on operating model design and RevOps strategy are useful when the same friction keeps surfacing between sales, delivery and support.

The WHO uses the same method at a very different scale. Its after action review guidance describes a qualitative review of actions taken, with the goal of identifying and documenting best practices. If the method is robust enough for emergency response, it is robust enough for a client onboarding.

FAQ

What are the four steps of an AAR?

The four steps are the four questions: what was supposed to happen, what actually happened, what were the differences, and what can we do differently next time. You work through them in order, record the answers, and finish with actions that have owners and dates.

How long should an after action review take?

For a single task or milestone, plan on 30 minutes to an hour, which is the range given in the NHS England handbook. Larger projects are better served by a longer lessons learned workshop or by several short AARs at each milestone.

Who should attend an AAR?

Everyone who took part in the activity, whatever their seniority, plus a facilitator who keeps the discussion on track. Leave out people who were not involved, since they add opinion without evidence. Share the written summary with them afterwards if they will run similar work.

Is an AAR the same as a lessons learned workshop?

No. An AAR is short, focused on one task and held soon after it. A lessons learned workshop is broader and usually covers a whole project with wider stakeholders. Many teams use AARs during a project and a workshop at the end.

What should we do if people will not speak openly?

Restate that the purpose is learning and that nothing said will feed into appraisals, ask the most senior person to speak last, and focus on processes rather than individuals. If silence persists, consider a neutral outside facilitator. Trust builds over repeated reviews where candour is not punished.