Lesson 20 of 23 · 7 min read
Many small experiments: daily Kaizen vs Kaizen events
In Lean: daily Kaizen and Kaizen events
Builds on: Every change is an experiment, Problems are not faults
In one minute
Teams improve work in two broad ways. A Kaizen event gathers people for a few focused days to redesign one process; daily Kaizen makes small changes part of every normal week. Events have their place, but it is the habit of many small experiments, each easy to undo, that keeps work improving after the event is over. A good experiment is one change, tried for about a week, small enough that undoing it does no harm.
The idea
A few focused days on one process (in Lean: Kaizen event). A group sets aside several days, follows the work, finds the causes and redesigns the process. It suits problems that cross several teams, need everyone in one room to agree, or need a large rearrangement that cannot be done a step at a time. Its weakness comes afterwards. Many changes land at once, everyone goes back to their usual work, and the checks and standards that make changes hold often never happen. With many changes at once, the checks also cannot tell which change helped.
Small changes as part of every week (in Lean: daily Kaizen). The people who do the work make one small change at a time, each a short PDCA cycle. Feedback is fast, risk is low, and the habit grows: a team that runs one small experiment a week has tried and checked a dozen ideas by the end of a quarter.
Designing a one-week experiment.
- One change, aimed at one cause.
- A small scope: one team, one kind of request, one form.
- A prediction, written before (lesson 14).
- A way back: how you will undo it, and how quickly.
- Agreement from the people it affects: whoever receives the work, and the owner of any control it touches.
- A date to look at the result.
- Plan. State the problem and its cause, choose the change, and write down what you expect to see.
- Do. Make the change on a small scale: one team, one form, one kind of request.
- Check. Compare what happened with the prediction, in the work itself, at set dates.
- Act. Held: make it the standard. Partly: adjust and go again. Reverted: back to the cause.
Why about a week? Long enough to see the change meet a run of ordinary days, short enough that a failure costs little. A good week is not the end of checking: if it looks right, the change continues with the usual checks at about two weeks and a month.
An experiment that cannot be undone is a decision, and deserves more care. Never "experiment" by quietly dropping a legal or audit control; agree any change to a control with its owner first, as lesson 6 described.
Why this depends on lesson 19. Experiments fail; that is their job. If a failed experiment is treated as someone's mistake, people stop running them, or run only safe ones that teach nothing. A team that counts "it did not work, and here is what we learned" as a good week runs more experiments, and learns faster.
Example
Illustrative example, not a real company.
A procurement team held a three-day workshop last year that produced a long list of changes. A few months later most had faded, and nobody could say which ones had helped.
This year they try a different rhythm. Week one, one change: purchase requests below a set amount, from one department, skip the department head's approval for a week. The finance controller agrees, and the department head will look at the list of those requests on Friday. Prediction: those requests are ordered within two working days, and the Friday list contains nothing she would have refused. Way back: switch the routing rule back, a five-minute job.
On Friday the prediction held. The change continues, with checks at two weeks and a month. The following week the team runs a separate small experiment on how quotes are requested, so each result can be read on its own.
Common mistakes
- Bundling changes. Five changes in one "experiment" make the check unreadable.
- No way back. If you cannot undo it, it is not an experiment.
- Saving problems for the next workshop. Most frictions can be tested this week.
Words to know
- Small changes as part of every week (in Lean: daily Kaizen).
- A few focused days on one process (in Lean: Kaizen event, sometimes called a Kaizen blitz).
- Way back: how an experiment will be undone if the check shows it did not work.
Check your understanding
Try it on your own work
Describe one recurring task that wastes your time, and see these ideas applied to it.
Where this comes from
Both focused improvement events and everyday small improvements are established practices in the Lean tradition that grew from the Toyota Production System; many Lean practitioners stress that the everyday habit is what makes improvement last.