TurningCircle
3 November 2026updated 16:06

Photograph: Pexels

Guides

The turning circle: why PDCA still drives continuous improvement

Plan, do, check, act is nearly a century old and still the most practical way to improve a process. What the loop actually is, where it came from, and how to run it without turning it into a paperwork exercise.

By The TurningCircle Team 3 min read

Most improvement programmes do not fail because the idea was wrong. They fail because the work was treated as a one-off project: a launch, a review, a report, and then a slow drift back to how things were. The PDCA cycle — plan, do, check, act — exists to prevent exactly that. It is a loop rather than a line, which is why it survives in factories, dealerships and software teams alike, and why it is still the default method taught for continuous improvement.

What PDCA actually is

PDCA is a four-step feedback loop for changing a process, a product or a service:

  1. Plan. Define the problem, find the root cause, and write down a hypothesis with a measurable target.
  2. Do. Run the change small enough that being wrong is cheap — one site, one shift, one cohort of customers.
  3. Check. Compare what happened with what you predicted. Not "did it feel better", but did the number move, by how much, and what surprised you.
  4. Act. Standardise what worked, discard or adjust what did not, and start the next cycle from the new baseline.

Then turn the wheel again. The repetition is the method, not a formality attached to it.

Where it came from

The loop is usually credited to Walter Shewhart, who described a specification–production–inspection cycle at Bell Labs in the 1930s, and it was popularised by W. Edwards Deming, who took it to post-war Japan where it became part of the foundation of what later became known as lean manufacturing and the Toyota Production System. Deming preferred the name PDSA — plan, do, study, act — to make the point that the third step is about learning rather than inspection. The two names describe the same loop; PDCA is simply the more common one in British management writing.

Why the circle beats the checklist

A checklist has an end. A circle does not, and that changes behaviour in three useful ways.

It makes uncertainty explicit. The plan step forces a prediction, and a prediction can be wrong. Teams that write down what they expect to happen learn far more from a failed test than teams that only record what they did.

It keeps the experiment small. Because the loop is expected to repeat, there is no pressure to make the first attempt definitive. A change that can be tested in a week beats a change that needs a quarter to evaluate, every time.

It separates learning from deciding. The check step is where evidence is gathered; the act step is where the decision is made. Merging the two is how organisations end up with confident conclusions and no data.

PDCA in practice

Take a familiar example from a franchised dealership: service department bookings are being lost because customers cannot get through on the phone at peak times.

  • Plan. Hypothesis: adding a second person on the phones between 08:00 and 10:00 will cut abandoned calls. Target: abandoned calls down by a third within four weeks.
  • Do. Run it for two weeks on Tuesdays to Thursdays only, and log every call.
  • Check. Abandoned calls fall by more than the target, but the average call length rises because the extra person is picking up the more complicated enquiries. The queue moved rather than disappeared.
  • Act. Keep the second line, and start a new cycle on the script used for the complicated enquiries.

Nothing in that sequence required new software or a consultant. It required a prediction, a measurement and the discipline to look at the result honestly.

Five ways to make the loop stick

  1. Start smaller than feels comfortable. If a test needs a quarter to produce a signal, it is too big to be a test.
  2. Write the hypothesis down. "We believe X will cause Y, measured by Z, by this date" is the whole discipline in one sentence.
  3. Check against the baseline, not against hope. Compare with the number before the change, not with the number you wanted.
  4. Standardise wins immediately. An improvement that is not written into the standard process quietly disappears within a month.
  5. Treat the loop, not the win, as the unit of progress. Teams that count completed cycles keep going through the ones that miss.

The bottom line

The turning circle endures because it matches how improvement actually happens: iteratively, with imperfect information, and in public. Plan carefully, test small, check honestly, act decisively — then turn the wheel again.