PDCA
Plan-Do-Check-Act
PDCA is the basic cycle of improvement: plan with a prediction, test on a small scale, compare result with prediction, and standardise or adjust. The cycle separates learning from “let’s try something”.
Use the method when
- →Any change whose effect is not certain in advance — i.e. most of them
- →As the engine in kaizen, A3 and experiments: all improvement formats run PDCA inside
- →When the organisation makes many efforts but harvests little learning
Don’t use it when
- →The change is trivial and risk-free — just do it (and check anyway)
- →As a rhetorical figure in presentations, while practice is Plan-Do-Plan-Do
Before you start
A baseline number exists: you cannot see an improvement without a before
The change can be tested in a bounded way — one line, one week, one case type — before rollout
How to do it — step by step
Plan — with an explicit prediction
1-2 hoursWhat is changed, where, over which period, and what do we expect to happen with which number? The prediction is what makes Check possible: “We expect the changeover time to fall from 45 to around 30 minutes.” Without a prediction, any outcome is “quite good”.
Do — on a small scale, with observation
days to weeksCarry out the change in a bounded way and be present: the deviations and surprises that arise along the way are data, not disturbances. Note them.
Check — result against prediction
1 hourMeasure the same number as the baseline and compare with the prediction — not just “did it get better?”, but “did it turn out as we expected — and if not, what did we misunderstand?” That question is where the learning lives.
Act — standardise or adjust
1-2 hours + rolloutDid it work as predicted: update the standard, train, and roll out — with the same measurement at the new places. Did it not work: adjust the hypothesis and run a new cycle. Both outcomes are progress; only the undocumented attempt is waste.
The classic mistakes
How it goes wrong in practice, and this is where most implementations part ways.
Plan-Do-Plan-Do
The organisation loves to plan and execute, and skips Check and Act, because the next initiative is waiting. The result is activity without accumulated learning: the same errors are made in new projects.
Cycles that are too big
A PDCA over six months gives one round of learning per half-year. Four cycles of six weeks give eight times as much learning in the same calendar time.
Check without a baseline
“It feels better” is not a Check. Without a before-number, the cycle is an anecdote.
Related standards
The standard is free. The anchoring is the craft.
The method only works when it becomes daily routines and leadership behaviour. The operations check shows in 4 minutes where your operations stand, and what should come first.