PDCA (Plan-Do-Check-Act) is a four-step improvement loop descended from Walter A. Shewhart's 1939 cycle and known as the Deming cycle; Deming himself preferred PDSA, Plan-Do-Study-Act. You plan a hypothesis and how to judge the result, test the change on a small scale, compare the result with the prediction, then standardise the change or loop back. Repeated endlessly, it is the engine under kaizen.

PDCA

Details
Full Name

Plan-Do-Check-Act

Also known as

the Deming cycle, the Deming wheel, the Shewhart cycle; PDSA (Plan-Do-Study-Act) in Deming’s own wording

Core Concepts:

Four Steps, Repeated Forever

The loop is deliberately small and deliberately endless:

  • Plan — state the problem, form a hypothesis about cause and remedy, and decide how the result will be judged

  • Do — carry out the change, preferably on a small scale

  • Check — compare the observed result with what was expected

  • Act — standardise the change if it worked, or return to Plan if it did not

The Scientific Method, Shrunk to Fit a Shop Floor

PDCA is an experiment loop: hypothesis, trial, evaluation, decision. Its ancestor is Shewhart’s 1939 sequence of specification, production and inspection, which he bent from a straight line into a circle precisely because "these three steps must go in a circle instead of in a straight line".

Small Scale First

The Do step is meant to be a test, not a rollout. Running the change on a small scale is what keeps a failed hypothesis cheap — and what makes the Check step interpretable at all.

Act Means Standardise or Loop

The step most often skipped. A successful change that is not written into the standard decays; a failed change that is not fed back into Plan teaches nothing. Act is where the loop either compounds knowledge or leaks it.

The Engine Under Kaizen

PDCA, the seven basic quality tools and the QC-story format became the foundation of kaizen in Japan. When a text says "continuous improvement" without naming a method, PDCA is usually what it means.

Ancestor of the Phase-Gated Cycles

dmaic is PDCA with gates, measurement discipline and statistics bolted on; dmadv is the same lineage applied to design. Both descend from this four-step loop.

Deming Preferred "Study"

Deming warned that "check" is misleading, because the English word means "to hold back". His own late formulation is PDSA — Plan-Do-Study-Act — where Study asks what was learned, not merely whether the target was hit. See Current Status: the naming split is not cosmetic.

When to Use:

  • Any improvement where the remedy is a hypothesis rather than a known fix

  • Small, reversible changes that can be tested before they are committed — the loop’s economics depend on the test being cheap

  • As the frame for retrospectives, incident follow-ups and process experiments that would otherwise end at "we should do better"

  • Where an organisation needs a shared vocabulary for "we tried it, here is what we learned" — PDCA gives the four slots

  • As the structural skeleton of a management system: ISO 9001:2015 and its siblings are organised around exactly these four steps

Common Misunderstandings:

  • ❌ "Deming invented PDCA" — he did not. Japanese executives recast his 1950 Deming wheel into PDCA in 1951; Deming later wrote, "How the PDCA ever came into existence I know not."

  • ❌ "PDCA and PDSA are two names for one thing" — Deming himself denied it: asked how the QC circle and the Deming circle relate, he answered, "They bear no relation to each other."

  • ❌ "Check means verify the deliverable" — Check compares result with prediction. Without a prediction recorded in Plan, there is nothing to check against.

  • ❌ "One pass through the cycle is a PDCA" — the method is the repetition. A single loop is an experiment, not an improvement system.

  • ✓ "PDCA is the scientific method wearing work clothes" — fair, and the reason its critics measure it against scientific standards.

Criticism:

  • Almost nobody runs it as specified. Michael J. Taylor and colleagues reviewed 73 peer-reviewed healthcare articles that claimed to apply PDSA and found that only two met all five core principles of the method. "Less than 20% (14/73) fully documented the application of a sequence of iterative cycles", and "only 15% (7/47) reported the use of quantitative data at monthly or more frequent data intervals to inform progression of cycles." Their conclusion is blunt: "the varied practice compromises its effectiveness as a method for improvement." Taylor, McNicholas, Nicolay, Darzi, Bell & Reed, "Systematic review of the application of the plan–do–study–act method to improve quality in healthcare", BMJ Quality & Safety 23(4), 2014, pp. 290–298. The finding cuts both ways: it is a criticism of practice, and it makes claims of PDCA’s effectiveness hard to evaluate, since what was measured was rarely the method.

  • Deming rejected the version everyone uses. Ronald D. Moen and Clifford L. Norman document the objection from the source. In a letter of 17 November 1990, Deming wrote: "Be sure to call it PDSA, not the corruption PDCA." He "warned his audiences that the PDCA version is frequently inaccurate because the English word 'check' means 'to hold back'". Moen & Norman, "Circling Back", Quality Progress, November 2010, ASQ. Calling PDCA "the Deming cycle" attributes to Deming a cycle he publicly disowned.

  • The two cycles are aimed at different levels. At a US Government Accounting Office roundtable, Deming was asked how the QC circle and the Deming circle relate: "They bear no relation to each other. The Deming circle is a quality control program. It is a plan for management. … Maybe you could say that the Deming circle is for management, and the QC circle is for a group of people that work on faults encountered at the local level" (same source). Teams that adopt PDCA for strategy are using a local fault-fixing loop for a job it was not shaped for.

  • Check without prediction degrades into review. The gap Moen, Nolan and Provost closed in 1991 by requiring prediction and theory in the Plan step is exactly the gap the Taylor review measured twenty years later: cycles run without recorded predictions and without data over time cannot produce the deductive-inductive interplay that makes the loop scientific.

Current Status:

  • The naming split is documented, and the popular name is the "corrupted" one. Moen and Norman lay out the lineage: Shewhart’s cycle (1939) → Deming wheel (1950, JUSE seminar: design, production, sales, research) → Japanese QC’s PDCA (1951, per Masaaki Imai, with no individual ever claiming authorship) → Ishikawa’s expanded PDCA for TQC (1985) → Deming’s Shewhart cycle (1986) → Deming’s PDSA, "the Shewhart Cycle for Learning and Improvement" (1993). PDCA won the popularity contest; PDSA carries Deming’s intent.

  • The split is now regional and professional, not resolved. Healthcare quality improvement writes PDSA, following Deming and the Model for Improvement that Langley, Nolan and colleagues built on the PDSA cycle in 1994. Manufacturing, Lean and the ISO management-system world write PDCA. Both priors are strong; a prompt that needs the learning emphasis should say PDSA and say why.

  • PDCA is embedded in the ISO management-system standards. ISO 9001:2015 names the Plan-Do-Check-Act cycle explicitly in its introduction and organises its requirement clauses along it — planning (6) as Plan, support and operation (7, 8) as Do, performance evaluation (9) as Check, improvement (10) as Act. Overview of the ISO 9001:2015 PDCA structure. That embedding is why the four-letter form will not go away, whatever Deming wrote.