What Is the PDCA Cycle (Plan-Do-Check-Act)?

What is the PDCA cycle?

The PDCA cycle (Plan-Do-Check-Act) is a four-stage method for making changes to a process, testing them on a small scale, checking whether they worked, and then either standardising or abandoning them. It's a structured way to run an experiment on your own operation instead of guessing.

Unlike a one-off fix, PDCA is meant to repeat. Each turn of the cycle either locks in an improvement or feeds a lesson into the next Plan stage — which is why it's usually drawn as a wheel or loop, not a straight line.

The four stages of PDCA

Stage	What happens	Typical output
Plan	Define the problem, the goal, and a specific change you believe will help. Predict what should happen if the theory is right.	A hypothesis, a small-scale test plan, success criteria
Do	Carry out the change on a limited, controlled scale — one line, one shift, one product run — not a full rollout.	Test data, observations, any deviations from plan
Check	Compare actual results against the prediction. Look for gaps between what you expected and what happened, and why.	Analysis, root cause of any gap, a verified conclusion
Act	If the change worked, standardise it (update the SOP, train the crew, lock in the new setting). If it didn't, revise the theory and go again.	A new standard — or a new Plan

The Check stage is where PDCA earns its reputation as more rigorous than "try something and see." It forces a comparison against a prediction made in advance, not a comparison against a vague sense of "better."

Who invented PDCA? Deming, Shewhart, or someone else?

This is more contested than most summaries admit, and it's worth getting right.

Walter Shewhart, the statistician who built the foundations of statistical process control at Bell Labs, described the original three-step cycle — Specification, Production, Inspection — in his 1939 book Statistical Method from the Viewpoint of Quality Control. He framed it as the scientific method applied to production: hypothesis, experiment, and test of the hypothesis.

W. Edwards Deming built on Shewhart's work. At a 1950 seminar for the Japanese Union of Scientists and Engineers (JUSE), Deming presented an expanded four-step version — design, produce, sell, and study the results via market research and redesign. Japanese engineers who heard it began calling it the "Deming wheel" or "Deming circle."

The PDCA label itself — Plan-Do-Check-Act, as commonly taught today — was not coined by Deming. According to Ronald Moen's history of the cycle published by The Deming Institute, "Japanese executives recast the Deming wheel presented in the 1950 JUSE seminars into the Plan-Do-Check-Act (PDCA) cycle" by around 1951, and "no one person claimed authorship" of that specific reformulation.

Deming himself objected to the PDCA name later in life. In a 1990 letter he wrote, "Be sure to call it PDSA, not the corruption PDCA," and in 1993 he formally proposed renaming it the "Shewhart Cycle for Learning and Improvement — the PDSA Cycle" — Plan-Do-Study-Act — arguing that "Check" implies holding back, while "Study" better captures the analytical work of that stage.

So the honest, sourced version is: Shewhart originated the underlying scientific-method cycle; Deming adapted and popularised it in Japan; unnamed Japanese practitioners produced the specific "PDCA" wording most of the internet now attributes to Deming alone; and Deming spent his later career trying to correct that attribution in favour of "PDSA." Both PDCA and PDSA describe the same basic loop — the difference is largely in the word for stage three, not the mechanics.

How do you use PDCA in manufacturing? A worked example

Take a common floor problem: a stamping line is running an 8% scrap rate on a bracket part, against a target of under 3%.

Plan. The team reviews scrap tags for four weeks and finds most rejects are cracked flanges, concentrated on the afternoon shift. Their theory: die lubricant is drying out mid-shift as the press heats up, increasing friction and cracking the metal on the bend. The plan: apply lubricant on a fixed 90-minute schedule instead of "when it looks dry," and track scrap rate and crack location for two weeks on one press only.

Do. One press runs the new lubrication schedule for two weeks. The other presses continue as normal, acting as a comparison. Operators log lubricant application times and any cracked-flange rejects by shift.

Check. Scrap on the test press drops from 8% to 3.4%; the untouched presses stay around 7-9%. Cracked-flange rejects on the test press are almost eliminated, but a small number of a different defect (dimensional drift) persist — worth its own investigation later. If the team wanted to understand why the original drying-out problem existed in the first place — a worn seal on the lubricant reservoir, say — that's exactly the kind of question a 5 Whys root cause analysis is built for, and it pairs naturally with the Check stage.

Act. The result confirms the theory. The 90-minute lubrication schedule becomes the new standard work for all presses, written into the operator SOP and the pre-start checklist, with a reminder timer added to each press. The dimensional-drift issue becomes the subject of the next PDCA cycle.

That's the whole method: a specific, falsifiable prediction; a small test; an honest comparison against the prediction; and a decision to standardise, discard, or iterate — not a company-wide mandate rolled out on a hunch.

How does PDCA relate to Kaizen and continuous improvement?

PDCA is the engine inside most continuous improvement systems, not a separate methodology competing with them.

Kaizen — the Japanese term for continuous, incremental improvement, popularised by Masaaki Imai — is the philosophy that improvement should be constant and everyone's job. PDCA is the mechanism that makes kaizen more than a slogan: it's the repeatable loop a team actually runs each time it wants to test an idea, whether that's a five-minute point kaizen on a workstation or a multi-month project.

Lean manufacturing, Six Sigma's DMAIC (Define-Measure-Analyse-Improve-Control), and ISO 9001's approach to corrective action all use a version of the same underlying logic as PDCA: define the problem, test a change at small scale, verify against evidence, then lock it in. PDCA is usually the simplest and fastest of these to teach a shop floor team, which is why it tends to be the entry point.

The common failure mode isn't choosing the "wrong" one of these — it's skipping Check. Teams plan and do constantly; fewer genuinely check results against a prediction before declaring victory and moving on. A PDCA cycle without a rigorous Check stage is just a change with extra paperwork.

FAQ

Is PDCA the same as PDSA? Functionally, yes — both describe the same four-stage improvement loop. PDSA (Plan-Do-Study-Act) is the version W. Edwards Deming preferred and considered more accurate; PDCA (Plan-Do-Check-Act) is the more common label today, coined by unnamed Japanese practitioners who adapted Deming's teaching in the early 1950s.

How long should one PDCA cycle take? There's no fixed rule — a cycle should be as short as the test allows. On a shop floor, a point-of-use PDCA might run for a single shift or a week; a process redesign might run for a full quarter. The discipline is in keeping each individual test small and time-boxed, not in a specific duration.

What's the difference between PDCA and Six Sigma's DMAIC? Both are structured improvement loops built on the same scientific-method logic. DMAIC (Define-Measure-Analyse-Improve-Control) is more statistically rigorous and typically used for complex, data-heavy problems with a dedicated project team. PDCA is lighter-weight and better suited to fast, iterative, shop-floor-level changes.

Do you need special training to use PDCA? No. PDCA is simple enough to teach in a toolbox talk, though getting real value from it takes discipline — mainly in writing down a genuine prediction before testing, and being honest in the Check stage when results don't match the plan.

Who actually invented PDCA — Deming or Shewhart? Neither, precisely. Walter Shewhart created the original three-step scientific-method cycle in 1939. Deming adapted and taught an expanded version in Japan in 1950. The specific "Plan-Do-Check-Act" wording was then coined by Japanese practitioners around 1951, with no single named author — and Deming later argued the whole thing should be called "PDSA" instead.

Ardmor works with Victorian manufacturers on both sides of this: the safety systems that keep a line compliant, and the lean/process discipline — PDCA included — that makes it actually run better. If you want an outside read on where your current processes and controls stand, book a process and safety audit.

This article is general information for manufacturers, not professional advice. Last updated: 20 September 2026.