PDCA Cycle Workshop

Purpose

The PDCA (Plan-Do-Check-Act) cycle, also known as the Deming cycle, is the fundamental engine of continuous improvement. This workshop guides the team through one full turn of the cycle: formulate a hypothesis, run an experiment, check results, and decide whether to adopt, adapt, or abandon.

At a glance

How to run

1. Plan (20 min) — Identify the problem or improvement opportunity. State it as a hypothesis: "We believe that [change] will result in [outcome] as measured by [metric]." Define the experiment: what will you change, for how long, and how will you measure?
2. Do (ongoing) — Execute the experiment. Keep changes small and contained. Document observations.
3. Check (20 min, after experiment) — Reconvene. Compare actual results against the hypothesis. Did the metric move? Were there unexpected side effects?
4. Act (20 min) — Decide: Adopt (make the change permanent), Adapt (tweak and run another cycle), or Abandon (the change did not work). Document the decision and rationale.

Materials needed

• PDCA template (hypothesis, experiment design, metrics, results, decision)
• Baseline data for the target metric
• Shared document for tracking the experiment

Pitfalls & common mistakes

• Skipping Plan — Jumping to Do without a hypothesis makes it impossible to Check meaningfully.
• Experiment too big — Keep experiments small and short so you learn fast.
• No Check — Without reviewing results, the cycle breaks and improvements are unverified.
• Always adopting — Not every experiment succeeds. Be willing to abandon.

Inclusion & accessibility

• Frame the experiment as learning, not pass/fail, to create psychological safety.
• Allow asynchronous input on the hypothesis for distributed teams.
• Document everything in writing so new team members can catch up.

Variations

• PDSA — Plan-Do-Study-Act: the Study phase is a deeper analysis than Check.
• Improvement Kata — Toyota Kata's structured approach to running PDCA experiments toward a target condition.
• Retrospective-driven PDCA — Use retrospective actions as the Plan phase input.

When NOT to use this

• If the solution is already known and proven — just implement it.
• If the team has no capacity for experiments — reduce WIP first.

Source & further reading

• Deming, W.E. (1986). *Out of the Crisis*. MIT Press.
• Rother, M. (2009). *Toyota Kata*. McGraw-Hill.