D-KAI-03 | First published 2026-07-18 | Revised 2026-07-18 | Primary sources: 3
Disclosure: Gemba Works Lab is supported by advertising. This page contains no affiliate links; if any are added in the future they will be clearly marked (per the FTC Endorsement Guides).
A kaizen event (also called a kaizen blitz) is a focused week in which a small cross-functional team fixes one bounded problem — one changeover, one cell, one chronic defect. Done well, it is the fastest way for a small plant to see what improvement methods can do. Done badly, it becomes "that week we rearranged things," and the area drifts back within a quarter. This checklist is written for a plant manager or supervisor running their first event in a shop of roughly 10–200 people, and it treats the follow-up period as part of the event, because that is where most events actually fail.
KEY POINTS
The event is 2 weeks of preparation + 5 days of execution + 30 days of follow-up. Skipping the first or last phase is the most common failure.
Scope one problem you can walk around in 10 minutes. First events die of ambition, not of difficulty.
Include the operators who work the area — both because it works better (a principle OSHA documents for safety programs) and because they keep the standard alive afterward.
The shape of the week
Figure: The four phases of a first kaizen event. The gain is created in phase 3 and kept in phase 4.Source: Gemba Works editorial team, structure aligned with NIST MEP continuous-improvement guidance.
If you want facilitation help rather than a DIY run, the NIST Manufacturing Extension Partnership operates a national network of centers whose advisors coach exactly this kind of event for small and mid-size manufacturers (source: NIST MEP — Improve). Documented MEP kaizen engagements show what a facilitated event can reach — one published example is covered in our kaizen patterns guide.
Preparation checklist (2 weeks out)
Before you commit the week
Pick one bounded target — one changeover, one cell, one defect. If you cannot state the problem in one sentence with a number in it ("changeover on press 2 takes 90 minutes"), the scope is too big.
Get the baseline number yourself. Time it, count it, or video it. An event without a baseline cannot prove anything on Friday.
Name a team of 4–7: at least two operators from the area, one person from a neighboring process, one skeptic. Involving the people who do the work is not a courtesy — worker participation is a documented success factor for finding and fixing workplace problems (source: OSHA — Worker Participation).
Arrange coverage. Team members are off their regular duties for the week. Half-committed teams produce half-events.
Pre-stage supplies: tape, cardboard, markers, stopwatch, a camera, spare shelving. Most first-event changes are built from exactly this list.
Agree the safety boundary: which changes need engineering/maintenance sign-off (guarding, electrical, anything load-bearing) before they are made — improvement never overrides lockout or machine-guarding rules.
Day-by-day checklist (the 5 days)
Day 1 — See the current state
Walk the process end to end; the whole team observes a full cycle.
Record reality: times, distances walked, searches, waits. No fixing today — day-1 fixes are guesses wearing work gloves.
Day 2 — Find the causes
Sort observations: which steps add value, which are walking/searching/waiting/adjusting?
Ask "why" until you hit something physical you can change this week.
Choose 3–5 changes; sketch them on paper before touching anything.
Day 3 — Make the changes
Build in cheap materials first (cardboard shadow boards, taped floor markings). Permanent versions come after the idea survives.
Anything inside the pre-agreed safety boundary goes through sign-off — no exceptions during event excitement.
Day 4 — Test and adjust
Run the process with the changes; measure against the baseline.
Let the operators break the new method and fix what annoys them — objections today are free; objections next month are drift.
Day 5 — Standardize and present
Write the new one-page standard with the operators, photos included.
Present results to management: baseline vs new number, and the 30-day list of leftover items with owners and dates.
Follow-up checklist (the 30 days that decide everything)
In Japanese plants, the week is the start of the work: what makes gains permanent is that the new method becomes the audited standard. That habit — not the event itself — is the part most worth importing, as explained in what kaizen actually means.
After the event
Week 1: Supervisor checks the new standard daily; every deviation is a question ("what's blocking the new method?"), not an accusation.
Week 2: Close at least half of the leftover-item list; convert surviving cardboard fixtures to durable versions.
Week 4: Re-measure the Friday number. Publish it next to the baseline — including any slippage, honestly.
Capture the know-how: if the improved method depends on one veteran's touch or judgment, record it now — our 5-step video SOP method takes about a day.
Schedule the next event only after the first standard has held for 30 days. A stack of decayed events is worse than none: it teaches the crew that nothing sticks.
Expectation-setting, honestly: a first DIY event typically produces a solid improvement on one bounded process and a trained team — not a plant transformation. That is a good trade for one week.
Note the difference from routine supervision: this event checklist runs once, for one intensive week. For the checklist you use every day or every week afterward — to catch small problems before they need another event — see our gemba walk checklist.
Last updated: 2026-07-18 | Primary sources referenced: 3 | This checklist is general guidance, not safety or compliance advice — machine guarding, lockout and electrical changes always follow your own procedures and OSHA requirements. Spotted an error? Tell us via the contact page.