The OODA Loop: A Decision Log and Comparison with PDCA
May 3, 2026
Author: Shusaku Yosa
The OODA loop—Observe, Orient, Decide, Act—is a way to update decisions as information changes. Associated with John Boyd, it emphasizes understanding the situation, not simply taking action faster.
It is not a one-pass checklist. Results feed back into observation, and new evidence can change the interpretation. This article provides an editorial adaptation for everyday marketing decisions.
Record something useful at each stage
Stage | Work | Record |
|---|---|---|
Observe | Gather what can be seen | Date, scope, counts and change history |
Orient | Interpret context and alternatives | Denominators, seasonality, measurement issues |
Decide | Select a next step | Why to change, investigate or wait |
Act | Execute within a defined boundary | Owner, deadline, impact and reversal condition |
The U.S. Air Force’s explanation highlights orientation’s importance. Look for evidence that contradicts the preferred explanation instead of selecting only observations that fit past experience.
A fictional form-completion problem
Suppose resource-request completions decline. Before changing the advertising budget or button color, examine entry sessions, form starts, successful submissions, devices and recent releases.
Stage | Illustration |
|---|---|
Observe | Entry volume is similar, but mobile successes decline |
Orient | Consider a new required field, device issue or broken success event |
Decide | Reproduce the issue through an authorized test |
Act | Fix the confirmed cause and test success, failure and reload behavior |
If only the measurement event failed, real submissions may be unchanged. If the form genuinely fails, there is no need to wait for a larger statistical gap before repairing it. Separate observed evidence from interpretation.
Compare OODA and PDCA
Perspective | OODA | PDCA |
|---|---|---|
Focus | Update interpretation and choice | Execute, check and improve a planned process |
Useful question | What should we do in this situation? | How should this process improve? |
Risk | Confusing speed with sound judgment | Continuing a plan after its premise changes |
Combination | Detect change and choose a test | Incorporate learning into repeatable operations |
OODA is not always fast and PDCA is not inherently slow. Both require evidence and review. Choose according to the decision rather than declaring one framework universally superior.
Include the decision not to change
Two weak days of advertising results may reflect purchase lag or reporting delay. Learning dynamics and concurrent changes also matter. A fixed rule to replace every underperforming ad after 48 hours ignores those conditions.
Condition | Example response |
|---|---|
Confirmed implementation failure | Repair or pause within authorized scope |
Small, noisy sample | Set observation conditions and a review date |
Several simultaneous changes | Narrow the investigation |
Agreed cost or customer-impact boundary breached | Follow the defined stop rule |
Missing decisive evidence | Make investigation the next action |
Delegated decisions still need budget limits, scope and escalation conditions. Speed does not justify breaking customer commitments or ignoring suppression preferences.
Keep a decision log
Record the time, question, observed evidence, alternative explanations, working hypothesis, options, chosen action and reason, owner, deadline, expected change, waiting or stop conditions, result and next observation.
The record can be brief. It should explain why the decision made sense with the information available at the time. Use it to improve observation, interpretation and action boundaries—not to judge a decision solely by whether its outcome happened to be favorable.
