Most attempts to improve something at work fail in roughly the same way. Someone has a good idea. The idea gets rolled out across the whole team. It produces some benefit and some new problems. Nobody is quite sure which outweighs the other, because there’s no clear way to measure either, and there isn’t really a way to roll it back. Six months later the team has quietly absorbed both the good and the bad of the new way, and the next good idea arrives.
The PDSA cycle is the alternative.
What it stands for
PDSA stands for Plan, Do, Study, Act. It comes from the work of W. Edwards Deming, who adapted it from earlier work by Walter Shewhart in the 1930s. Deming used it for the rest of his career as the spine of how he taught improvement — first to Japanese manufacturing in the 1950s, then to the rest of the world.
The four steps look almost embarrassingly simple. The discipline is in actually doing them.
Plan
Identify a small, specific change you want to try. Predict what will happen if you make it.
Not what you hope will happen. What you actually predict will happen. The prediction is the important part. Without it, you’ll learn nothing — because whatever happens, you’ll be able to interpret it as confirmation of what you already believed.
If your prediction matches the result, you’ve confirmed something you already knew. If it doesn’t match, you’ve learned something new. The mismatch is where the value is.
The change should be small. Small enough to try in a week or two. Small enough that if it goes badly, you can stop without damage. Most “improvement” projects fail because they’re too big to actually test.
Do
Make the change. Run it for the planned period. Watch what happens.
The watching is the part most teams skip. They make the change, get busy, and stop paying attention until someone complains or something breaks. By then it’s too late to see the early signals — the small things that would have told you, in week one, whether the change was working or not.
Study
Compare what actually happened to what you predicted. Be honest.
This is where defensive routines kick in. If your prediction was wrong, the temptation is to revise the prediction in hindsight — “well, what I really meant was…” — so the result still looks like a success. Deming’s word here was study, not check, deliberately, because checking is what you do when you’re hoping to confirm a result. Studying is what you do when you’re actually trying to learn.
What did you predict? What happened? Where do they differ? What does the difference tell you?
Act
Based on what you learned, decide what to do next.
There are three honest possibilities. Adopt the change, because the result was clearly better than the alternative. Abandon it, because the result was clearly worse. Adapt it, because the result was mixed and there’s a smaller change worth trying next.
Most teams treat this as a forced choice between adopt and abandon. The third option — adapt and run another cycle — is usually the most useful.
Then you start again.
What makes PDSA different from “just trying things”
Three things.
It’s small. A PDSA cycle should fit in a week or two. The point isn’t to make the change permanent — it’s to learn whether the change is worth making permanent.
It’s predicted. The prediction is what turns “trying things” into actual learning. Without a prediction, you can’t tell whether you were right, you can’t tell whether you were wrong, and you can’t update your understanding of how the work actually works.
It’s a cycle. The output of one PDSA is the input to the next. A team that runs PDSA cycles consistently doesn’t make one big improvement a year — they make many small improvements continuously, each one informed by what they learned in the last.
A small example
You think your team meetings are running too long.
Plan. You decide to try ending the meeting five minutes earlier next week and see what happens. Your prediction: people will be more focused because they know the meeting will be shorter, and the meeting will end on time without missing anything important.
Do. You run the meeting with a hard 25-minute limit instead of 30.
Study. What actually happened? Did the meeting feel rushed? Did something important not get covered? Did people leave the meeting clearer or more confused?
Act. Based on what you learned, either keep the shorter format, go back to 30 minutes, or try 25 minutes with a different agenda structure.
This sounds trivial. It is. That’s the point. Big improvements are rarely the result of one big change. They’re the result of many small, well-tested ones.
Where this fits
The PDSA cycle is one piece of Deming’s broader System of Profound Knowledge — his framework for what it actually takes to lead an organisation that improves over time.
It’s also a practical way to do what double-loop learning calls for: not just adjusting how you do things, but questioning whether you’re doing the right things at all.
Sources
The cycle was originated by Walter Shewhart in the 1930s and developed extensively by W. Edwards Deming through the 1950s onwards, particularly in his books Out of the Crisis (1982) and The New Economics (1994).