Plan-Do-Check-Act

Plan-Do-Check-Act is the engine of every lean shop's improvement habit. The cycle is older than lean itself, traceable to Walter Shewhart in the 1930s and brought into manufacturing by W. Edwards Deming and Toyota over the following decades. The structure is deceptively simple: think before you change, change at small scale, look at what actually happened, then decide what to do next. The discipline comes from running the loop quickly and continuously, not from any single pass through it.

"Small loops, fast loops, many loops. The improvement is not in the change, it is in the rhythm of changing."

How PDCA works

PDCA is a four-step cycle:

1. Plan. Write down the problem in specific terms. Form a hypothesis about the cause. Predict what will change if the countermeasure works. The plan is not a long document; it is a few sentences. A one-page A3 is the format many shops use to capture it.
2. Do. Run the change as a small pilot. One shift, one machine, one product family. The Do phase is not the full rollout; it is the experiment.
3. Check. Measure actual results against the prediction. Look for both the gains and the side effects. A change that solves the target problem but creates a new one downstream has to be caught in Check.
4. Act. Decide. If the change worked, standardize it: update standard work, update training, update the schedule. If it did not, return to Plan with what you learned and a refined hypothesis.

The cycle is meant to be fast. Most PDCA loops on a shop floor are two to four weeks long. The discipline that distinguishes a lean shop from a firefighting shop is not the depth of any single PDCA loop. It is the cadence of running loops back to back, one after another, on every problem the team can see.

PDCA is iterative by design. The Act step is not the end. It loops back into the next Plan, with a new problem or a deeper version of the same one. Over months and years, the compound effect is large; in a single quarter, no single loop looks like much.

Where PDCA fits on the shop floor of a small manufacturer

Imagine a 30-person plastics injection molding shop where finished parts have been piling up in the staging area at the end of the day. The owner has been blaming the packing line. Before assuming, the shift lead runs PDCA.

Plan, week one. The problem statement is specific: finished parts average eight hours of staging before reaching the packing line, and the staging area runs out of floor space twice a week. The hypothesis is that the press operators are running full bins before sending to packing rather than sending half-full bins on a fixed interval. The prediction: if bins are sent every two hours regardless of fill level, staging time will drop to under three hours and floor-space crises will stop.

Do, week two. The team tries the new rule on the day shift's two presses only. Check, end of week two. Actual staging time drops to two hours. Floor-space crises stop on the trial shift. The packing line absorbs the smaller batches without trouble. One side effect: a slight increase in cycle time at the presses because of more frequent handoffs, which the team decides is acceptable.

Act, week three. The new rule is standardized across all shifts and presses. The lead writes a one-paragraph standard-work update. The next PDCA loop targets the next biggest staging issue: kit pulls that arrive late at the assembly station.

That is PDCA at small scale. No software, no consultant, a two-week loop, a problem solved, a new loop started.

Common mistakes with PDCA

• Skipping Plan and going straight to Do. A change without a written hypothesis and prediction cannot be checked.
• Treating Do as the full rollout. The Do phase is a pilot, not a launch. Failures at pilot scale are cheap. Failures at full rollout are not.
• Skipping Check. Many shops install a change and never look at whether it worked. The change either fades or causes a downstream problem that goes unnoticed.
• One-and-done PDCA. A single loop is not lean. Lean is the rhythm of running loops continuously. The Act step has to lead into the next Plan.
• Confusing PDCA with a project plan. A project plan has a finish line. PDCA does not. Treating the cycle as a one-time event misses the point.

PDCA and related Lean tools

PDCA is often captured on a one-page A3 document, which makes the cycle visible and reviewable. The Deming variant of the cycle is Plan-Do-Study-Act, with Study replacing Check. The Six Sigma counterpart is DMAIC, a heavier, more statistical framework for variation problems. PDCA is the iterative engine behind kaizen, the continuous improvement habit that keeps a lean shop from sliding back.