PDCA Review Sheet
Connect a plan, recorded execution, a comparable check and a conditional next action in one review sheet.
Use a bounded cycle to compare a work trial with its original plan, including quality failures. Keep the four columns and the underlying records so the next cycle can test what remains uncertain.
Cycle scope, comparable baseline, hypothesis, target and quality guardrail, execution records, denominators and exceptions.
An evidence-based P/D/C/A review, remaining explanations and a conditional next-cycle plan.
On this page
How to use it
- Define one cycle, baseline, measurement rule and quality guardrail.
- Record the planned hypothesis and the actual actions separately.
- Recalculate comparable results, then inspect quality and alternative explanations.
- Choose a conditional next action with owner, resources and approval status.
- Start a new plan without overwriting the previous record.
Keep its limits in mind
Separate plan, execution, interpretation and next action. Keep counts, percentage points and relative changes distinct. A before/after sample does not establish causation; a speed target does not override failed quality. This working sheet is not a certification checklist. Proposed standardization needs a decision; PDCA and PDSA terminology remains distinct.
A speed target passes while quality fails
Original fictional I5 teaching packet: Clearbrook Support. Records, statements, trials and agreements are synthetic, not real outcomes or personnel assessments.
On a small screen, scroll across the diagram to read it.
Recalculate both conditions: The baseline has 32 timely requests and the trial 42 out of 50 each. The speed target passes, but two missing owners fail the zero-missing guardrail.
Decide what can follow: Uncontrolled periods cannot establish causation. Keep the original cycle and propose a field check and load log before another trial; implementation and standardization remain unapproved.
Keep the extra10 timely requests, +20 pp and+31.25% relative distinct; two quality failures still matter.
Next step: If the deviation is recorded but load and form differences remain unclear, inspect resources by session and practice before choosing a change.
Three reasons to pause before standardizing a trial
Use PDCA to retain what you planned, what happened, what the evidence supports and what decision should follow. State both the desired result and the conditions that must not be sacrificed. A cycle is useful even when the answer is to revise or stop, provided the underlying records remain available.
Pause a standardization decision when a quality guardrail fails, even if the speed target passes. Pause comparison when the denominator or meaning of “timely” changed between periods. Pause a causal claim when workload, staffing or other changes were not controlled. You can still describe the observed difference while acknowledging that its cause is unresolved.
Agree the scope, measurement and decision rights before the trial. Avoid turning Check into a search for flattering numbers or Act into an automatic rollout. If the next action needs approval, label it as a proposal and keep the original cycle separate from any future cycle.
Choose a cycle, a reflection or a redesign
PDCA organizes a bounded trial and its follow-up. PDSA uses Study in its name; keep that terminology when working in a PDSA process rather than silently renaming it. Neither acronym replaces an actual record of the experiment.
| Method | Input | Output |
|---|---|---|
| PDCA | A plan, execution records and agreed checks | A comparison and proposed next action |
| PDSA | A planned test and learning questions | Study and adaptation in that named cycle |
| KPT | A bounded period and attributed observations | Keep, Problem and Try items |
| ECRS | A mapped process and required controls | Redesign candidates before a trial |
Worked case: recalculate speed without removing the quality failures
This continues fictional Clearbrook packet I5. The baseline is September 21–25; the trial is September 28–October 2. Each period contains 50 resolved requests, ten per day. “Timely” means before the next scheduled handoff. The hypothesis is that clearer instructions help; the targets are at least 80% timely and zero missing owners. These are specified teaching record days, not an inferred holiday calendar; they do not cover every incoming or unresolved request.
The daily timely counts are reproduced below. Baseline totals 32/50 = 64%; trial totals 42/50 = 84%. That is ten additional timely requests, a 20-percentage-point increase and a relative increase of 20/64 = 31.25%. Those three expressions describe different units and must not be interchanged.
All 50 trial records were also checked for owners, and two were missing one: 2/50 = 4%. This fails the zero-missing condition. The material does not say whether those two were timely or whether baseline owner completeness was checked. Do not subtract them from 42 or invent a baseline quality improvement. The speed denominator remains 50; the second condition is assessed separately.
| Day | Baseline timely / 10 | Trial timely / 10 |
|---|---|---|
| 1 | 6 | 8 |
| 2 | 7 | 8 |
| 3 | 6 | 9 |
| 4 | 7 | 8 |
| 5 | 6 | 9 |
| Total / 50 | 32 | 42 |
Conclude on both conditions, then plan the next check
The Check conclusion is that speed passes its target while quality fails. It is not overall success. Uncontrolled before-and-after periods cannot establish that clearer instructions caused the difference. Workload and form differences remain unclear, so preserve those questions alongside the improvement rather than treating them as footnotes.
The coordinator proposes an owner-field check and a load log before a limited extension. No extension or standardization has been approved. The request should identify the change, who may authorize it, what records would be kept and when to stop or review. A proposed mandatory field may alter how work is done; confirm the operating rule and handling of exceptions instead of assuming a form change fixes ownership.
If another trial is authorized, create a new cycle with its own dates, scope and actual approval. Keep the original I5 counts intact. Record request load, staffing availability, form version, unresolved work, timely status and owner checks under agreed definitions. Review both speed and quality before deciding whether to continue, adjust or stop. Distinguish implementing the owner check from proving it reduced missing owners; then distinguish an observed change from a causal explanation. Retain negative findings so the next cycle learns from the failed guardrail.
Check your work
- Cycle and four kinds of information are explicit.
- Rates and quality checks can be traced to records.
- A proposal is not approved standardization or an executed next cycle.
Framework references
References explain the framework and its scope; they do not validate the fictional example's results.
Put the framework to work
Copy the outline and add your question, evidence, and open questions.