Plan-Do-Check-Act (PDCA)

Plan-Do-Check-Act (PDCA) (CSSGB BoK VI.B.5 — Apply)

Quick Answer: PDCA is an iterative improvement cycle: Plan the change and measures, Do a small-scale trial, Check (study) results against expectations, and Act to standardize success or revise the plan. Green Belts apply PDCA inside and after DMAIC—especially in Control—so learning becomes standard work rather than a one-time project event.

PDCA is also called the Deming cycle or Shewhart cycle. Some organizations use PDSA (Plan-Do-Study-Act); for CSSGB, treat Check/Study as the evaluation step that compares data to the plan’s prediction. VI.B.5 is Apply: place activities into the correct PDCA step and use the cycle to sustain and extend gains.

Why PDCA Matters in Control

DMAIC projects are finite. Processes are not. After the team reduces defects, the organization still faces:

• New failure modes and customer requirements
• Operator turnover and equipment wear
• Opportunities to further reduce waste and variation
• Signals from control charts that need structured response

PDCA provides a simple, repeatable loop that process owners can run without launching a full Black Belt project for every tweak. Control-phase deliverables (control plan, training, audits, visual management) work best when embedded in ongoing PDCA.

The Four Steps Distilled

1. Plan

Define what you will improve and how you will know it worked.

• Clarify the problem or opportunity (gap from standard or goal).
• Set a measurable objective and success criteria.
• Analyze likely causes enough to propose a change (use data, process knowledge, prior FMEA).
• Design the change: who, what, where, when, and how.
• Predict expected results and select metrics and data collection methods.
• Plan the scope of the trial (pilot area, sample size, duration, risks, rollback).

Planning without measures turns “Check” into opinion. Planning without a small-scale boundary turns “Do” into uncontrolled plant-wide risk.

2. Do

Implement the planned change on the planned scale, preferably a pilot.

• Train the people involved in the trial.
• Execute the new method, tool, or setting as designed.
• Collect the data specified in the Plan.
• Note unexpected obstacles, but avoid silent mid-trial redesign that confounds learning (unless safety requires stop).

“Do” is not “roll out to the enterprise on day one.” Large uncontrolled changes make it hard to learn what worked.

3. Check (Study)

Compare results to the plan’s predictions and success criteria.

• Analyze metrics: did the change move the CTQ as expected?
• Review side effects: cost, cycle time, safety, morale, downstream impact.
• Assess implementation fidelity: was the change actually followed?
• Capture lessons: what was confirmed, refuted, or newly discovered?

Check is evidence-based. Celebration without data, or blame without analysis, both break the cycle.

4. Act

Decide what permanent action follows from the Check.

If the change worked (and side effects are acceptable):

• Standardize the new method (update SOPs, control plan, visuals).
• Train the full affected population; update training plans and records.
• Expand the change beyond the pilot with appropriate controls.
• Establish ongoing monitoring so regression is detected.

If the change did not work or was incomplete:

• Adjust the plan based on what was learned.
• Abandon harmful ideas deliberately (document why).
• Run another PDCA cycle with a refined hypothesis—do not pretend success.

Act is where sustainment lives. Without Act, PDCA is only a pilot report.

PDCA Cycle Table (Exam Memory Aid)

PDCA vs DMAIC

They are complementary, not rivals:

• Improve often contains mini-PDCA pilots of candidate solutions.
• Control installs the system (control plan, documents, training, audits) that enables process owners to run PDCA routinely.
• When control charts signal special causes, reaction plans may trigger a focused PDCA to find and remove the cause.

A full DMAIC project can be viewed as a large, rigorous instance of learning cycles; PDCA is the portable loop that remains when the project team disbands.

Applying PDCA After Handover (Green Belt Scenarios)

Scenario A — Chart signal: Plan a hypothesis for the special cause; Do a temporary containment and trial fix; Check whether the signal stops and capability holds; Act by updating the reaction plan or standard if the fix is permanent.

Scenario B — Audit finding: Plan corrective action for missing revision at the station; Do the fix and brief the shift; Check a follow-up audit; Act by improving distribution of controlled documents.

Scenario C — Further improvement: Plan a small layout change to cut motion waste; Do it in one cell; Check cycle time and quality; Act by standardizing the layout across cells if successful.

Common Exam Traps

• Calling “train everyone plant-wide on day one” the Do step without a Plan or pilot—scope and prediction belong in Plan; uncontrolled big-bang rollout is poor PDCA practice.
• Stopping after a successful pilot without Act (no SOP/control-plan update)—gains will not stick.
• Treating Check as optional because “the team liked the idea.”
• Confusing PDCA Act (standardize or adjust) with theatrical activity without decision.
• Assuming PDCA replaces SPC or control plans—it organizes how you use those tools over time.

Linking PDCA to Other Control Tools

• Control plan: documents the standardized method that emerges from successful Act steps and the monitoring used in Check.
• Document control & training: formalize Act when a new method becomes the standard.
• Audits: provide Check evidence that the standard is followed.
• TPM and visual factory: create visible, reliable conditions so Plan and Check use trustworthy signals from the workplace.

Exam Focus

Apply-level questions present a story and ask which PDCA step is occurring, which step is missing, or how PDCA relates to DMAIC Control. Anchor each answer to the core question of that step: design and predict (Plan), trial and collect (Do), compare to prediction (Check), standardize or revise (Act).

Key takeaway: PDCA is the continuous loop that plans a change, trials it, studies the evidence, and acts to standardize success—keeping DMAIC Control alive long after the project closes.