Kaizen event: the 5 mistakes that ruin it
A badly prepared kaizen event isn't neutral. It creates problems where there weren't any and burns your credibility for the next one.
A kaizen event is a short, intense session — three to five days — where a team works exclusively on redesigning one concrete slice of the process.
Enter your email and you will get an alert when something new is published. Unsubscribe anytime.
Set up well, it changes something in a week that had been the same for years.
Set up badly it isn't harmless: it eats your people's time and burns the credibility of the next attempt. That last part is the expensive one.
These are the five mistakes that spoil it.
1 · Not defining the scope
The most frequent one, and the one that guarantees failure.
"Improve assembly" is not a scope. "Cut changeover time on line 3" is. Without a clear boundary the team scatters, and on Friday you have plenty of ideas and nothing implemented.
2 · Bringing the wrong team
A kaizen event without the people who work at that station produces solutions that don't survive Monday.
And the other way round: if there's nobody who can decide on spending, your improvements sit waiting for an approval that arrives in March.
3 · Not measuring first
If you don't know where you started with data, at the end you can't demonstrate the improvement. And what can't be demonstrated can't be defended the day somebody asks whether it was worth it.
4 · Finishing without implementing
The goal is that the change is running on the last day. Not that an action plan is left behind.
An event that produces a list of pending tasks is a long meeting with a Japanese name.
5 · No follow-up
Without a review at thirty days, the process goes back to how it was. It's the law of gravity of continuous improvement and it spares nobody.
You fight it with two things: a date in the calendar and an owner with a name.
| Mistake | Symptom | Antidote |
|---|---|---|
| Fuzzy scope | "improve the plant" | one process, one metric, one boundary |
| Wrong team | nobody from the station | operators + whoever decides |
| No baseline | "this feels better" | measure before touching |
| Not implemented | action plan | change running on the last day |
| No follow-up | three months on, identical | 30-day review with an owner |
The bridge to 2026
Of those five, two have changed in nature over recent years: measuring first and following up.
Setting up the baseline used to be part of the event's work. A stopwatch, sheets, and two days of the week gone on measuring instead of improving. When the line already records its own times and stoppages, you start on Monday with the data on the table and you have the whole week to redesign.
And the thirty-day follow-up stops depending on somebody remembering. If the metric measures itself, the relapse shows the moment it starts — not three months later, when everything's already back where it was.
Let me tell you where most of the improvements I've watched get lost actually get lost: not in the week of the event. People throw themselves into that. They get lost in the month afterwards, when nobody is looking any more.