Problem solving Deliver

OODA Loop

The OODA loop is observe, orient, decide, act: a fast cycle for decisions that have to be revised as the situation moves.

The questionWhat has changed since we last looked, and does our picture of the situation still fit?

Also called Boyd's loop or the Boyd cycle.

Process · John Boyd, 1970s to 1995

ObserveCollect signals, including ones that could contradict you
OrientWrite the working picture and check it against new facts
DecidePick a testable action with a time limit
ActDo it at a size where the effect shows quickly

Simplified four-step cycle. In Boyd's own sketch, orientation shapes the other steps and feedback returns to it.

When to use it

A client works in conditions that change faster than its planning cycle, such as incident response, a product launch against a reacting competitor, or a fast-moving negotiation.
Teams keep acting on a picture of the situation that was accurate a month ago.
Decisions wait on an approval chain while the facts change, and you need to move decisions to the people who see the facts.
A review after a miss shows the information was available, but the team's existing picture filtered it out.

When to avoid it

The problem is stable and can be analyzed. In Cynefin terms it is Complicated, and a method such as Kepner-Tregoe gets better decisions than a quick loop.
Nobody in the room has authority to act. A faster loop of recommendations is just faster reporting.
The decision needs a formal, documented, auditable process. The loop gives you no structure for that.
The ask is simply to move faster. A client with no particular loop in mind will take the vocabulary and change nothing.

Overview

The OODA loop describes decision making as a repeating cycle of observing, orienting, deciding, and acting, with the result of each action feeding the next observation. It supports decisions that have to be revised as conditions change, such as incident response, a competitive launch, or a fast-moving negotiation.

Boyd put the weight on orientation: the mix of culture, experience, new information, and analysis that decides what an observation means. Speed matters, but a fast loop with a stale picture only produces errors sooner. In use, most of the work is making the picture explicit and noticing when it no longer fits.

What shapes orientation

Culture and heritage

Boyd lists cultural traditions and genetic heritage. For an organization, read this as its norms, habits, and what it treats as normal. It decides what counts as a signal. An operations culture sees a late shipment as a failure, and a sales culture may see it as a customer problem.

Previous experience

What the people involved have seen before. It makes orientation fast, and it is also what makes teams misread a new situation as an old one. Ask which past incident the team thinks this resembles.

New information

What comes in from observation. It only changes the picture if the team lets it. Look at what the current picture would predict, and where the data disagrees.

Analysis and synthesis

Breaking the situation down and putting it back together, across several domains or sources. It is how a team builds a new picture when the old one fails. It takes time, so it cannot be done only under pressure.

What it is not

A speed contest. Boyd wrote about getting inside an opponent's loop, but his late briefing puts the weight on orientation and the ability to adapt. A fast loop with a wrong picture produces errors sooner.
Four boxes in a line. In Boyd's diagram orientation shapes what is observed, decided, and done, and feedback returns to it. Orient is not just a step between observe and decide.
A business planning method. Boyd built it for combat and the business versions came later, from people applying it. Its critics find it vague, and no standard business procedure exists.

The steps

Observe

Decide what you collect and from where, and include sources that could contradict your current view. Ask what you would expect to see if the current picture were wrong. The output is a short list of signals with owners.
Orient

Write the current picture in a few sentences: what is happening, why you think so, and which past cases you are matching it to. Compare it with new information and look for mismatches. The output is a written working picture with its assumptions marked.
Decide

Choose an action as a hypothesis, with a time limit and the signal that would show it worked. Say who may act without asking, and within what limits. A decision that cannot be tested is a bet.
Act

Do it, at a size where you can see the effect quickly. Record what was done and when.
Loop back

Compare results with the signal you named. When they differ, rewrite the picture and not just the action. Set a rhythm for how often the loop runs, tied to how fast the situation changes.

In practice

The retailer's checkout error rate had doubled by the time the on-call lead opened the dashboard. The team's picture was a payment-provider outage, because three of the last four incidents had been exactly that. A deployment had gone out twenty minutes earlier, though, and one engineer raised it. The lead decided to roll back, as a test, and told the team what result would show the deployment was not the cause.

The errors continued, so the picture changed, and the team looked at provider status by region and found the fault in one. The review afterward found that the slow part was orientation and not the rollback. The runbook assumed a provider fault, so nobody had written down the deployment hypothesis or the signal that would rule it out. The team rewrote the runbook to list the working hypothesis at each update.

What you leave with

A written working picture of the situation, with assumptions marked.
A decision phrased as a testable action, with a time window and the signal that it worked.
A statement of who may act without escalation, and within what limits.
A record of each loop, covering what was seen, what was believed, what was done, and what happened.

Failure points

Faster without better

The team shortens its reporting or approval time while the picture stays wrong. It reaches the wrong action sooner.

Orientation never written down

Each person holds a different picture. The loop runs separately in several heads, and the disagreement appears only as conflicting actions.

Decision rights unchanged

Observations reach people who cannot act on them. The loop is fast at the edge and slow at the center, so nothing changes.

Observation filtered by the old picture

Dashboards show only what the existing view counts. Boyd's diagram has orientation shaping observation, so the filter is built in, and evidence that would break the picture never gets collected.

Vocabulary without a loop

A client talks about getting inside a competitor's decision cycle but cannot say what the cycle is, who runs it, or how long it takes. The phrase becomes a slogan.

The failures above are usually Framing: the wrong problem was chosen before this framework was applied.

Origin

John Boyd was a US Air Force fighter pilot and strategist who developed the loop in the early 1970s from questions about air combat, and then applied it to maneuver warfare more broadly. He wrote little and taught through briefings that he kept revising. His essay Destruction and Creation is dated September 1976, Patterns of Conflict was circulated in 1986, and a late briefing, The Essence of Winning and Losing, carries the version of the diagram where orientation is the center of the model. Air University Press later published the briefings as A Discourse on Winning and Losing.

The four-box circle in most business slides is a simplification. Boyd's own sketch shows orientation shaping observation, decision, and action, with feedback and outside information flowing back in. Business applications came later, notably Chet Richards, a retired Air Force colonel who worked with Boyd, in Certain to Win (2004). Critics, such as the aviation historian Michael Hankins, argue that the loop is vague enough for defenders and attackers to see what they want in it.