PDCA: turn improvement plans into measurable results
Plan, test, check and adjust: a practical method for improving processes and demonstrating results, with an IT example and a working outline.
In 30 seconds
- Goal: achieve and verify measurable improvement.
- Cycle: Plan → Do → Check → Act, then repeat.
- Key point: decide using results, not just completed actions.
- Use: quality, processes and IT operations.
Understand the PDCA cycle
PDCA means Plan, Do, Check, Act. It helps teams test an improvement, assess its effects and decide what to do next. The cycle repeats using what has been learned.
Completing an action does not yet demonstrate improvement. The aim is to connect a specific problem, an action and an observable result. This guide offers a practical application to quality and IT operations.
1. PLAN — Define the improvement
Describe the problem without assuming its cause. Define the process, affected users and observation period. Establish a baseline using verifiable data, then agree an objective and deadline.
- Who owns the work and who makes decisions?
- Which measure will demonstrate the result?
- Which hypothesis will be tested?
- What risks, dependencies and stopping criteria apply?
Useful record: problem, baseline, objective, hypothesis, owner and test plan. A numerical target must state what is counted and within which scope.
2. DO — Run a controlled trial
Test the action within an appropriate scope. Inform the people involved, confirm the necessary approvals and plan recovery if the trial fails. Use your change process for technical modifications.
Record dates, actions actually completed, deviations and observations. A pilot should be representative enough to learn from while limiting the consequences of an unexpected outcome.
Useful record: trial log, observations and any incidents.
3. CHECK — Assess and understand
Compare the results with the baseline and objective using a consistent definition of the measure. Also examine unintended effects: work transferred to another team, extra delays or reduced quality.
Investigate differences. Fewer incidents may reflect lower activity or a change in reporting. Distinguish what the evidence shows from what remains a hypothesis.
Useful record: before-and-after comparison, user feedback, analytical limitations and an effectiveness conclusion.
4. ACT — Decide, adapt and sustain
Use the results to decide whether to expand, extend, change or stop the trial. If the action works, update procedures, responsibilities and training. Check that the result is sustained.
If the objective is not met, revise the hypothesis or plan. Repeating the cycle using what was learned is a useful decision, not a failure to hide.
Useful record: decision, rationale, next actions, owners and review date.
Example: reduce recurring IT incidents
Fictional example: these figures illustrate the method and are not a performance promise.
Plan: over eight weeks, 20 recurring access incidents are recorded across 200 comparable requests. The team targets no more than 6 incidents per 100 requests, down from 10, without increasing processing time. The hypothesis is that a standard access check during employee onboarding will reduce rework.
Do: trial a checklist in one team for a defined period, train participants and record deviations.
Check: 10 incidents occur across 200 requests, or 5 per 100. Before drawing a conclusion, the team assesses request comparability, processing times, extra checklist effort and user feedback.
Act: if the findings hold, refine the checklist, expand gradually and maintain a review. Otherwise, investigate other causes and adjust the pilot.
For event classification, see the IT incident triage guide.
A simple outline for your next cycle
- Problem, scope and baseline.
- Objective, measures and success criteria.
- Hypothesis, actions, risks and responsibilities.
- Trial results and observed deviations.
- Effectiveness assessment and unintended effects.
- Decision, standardisation or another trial, then review date.
A shared document or existing tool may be enough. Match access controls, evidence retention and approval requirements to your organisation’s risks and obligations.
Common mistakes to avoid
Stopping at Do, choosing measures after acting, confusing activity with results, expanding too soon or leaving decision ownership unclear all weaken the cycle. PDCA provides a framework; it does not replace root cause analysis or the judgement of accountable people.
Read the PDCA glossary definition.
Method reference: ASQ — Plan–Do–Check–Act cycle. The IT example and working outline above are practical NetQualIT proposals.
Deliverables
- A four-step approach.
- A cycle outline to reuse in your tracking tool.
- A numerical example distinguishing activity from effectiveness.
Editable Excel workbook with FR and EN tabs. Record evidence, owners, deadlines, actions and decisions.
Operational summary
Define the intended result, test within a controlled scope, compare evidence and decide. The next cycle starts with what the previous one taught you.
Download
Structure your continuous improvement approach
NetQualIT helps you define objectives, select useful measures and establish reviews proportionate to your risks.