Kaizen Is Not an Event

Somewhere along the way, “kaizen” became “kaizen event” — a structured 3-5 day workshop with a charter, a facilitator, daily report-outs, and a Friday celebration. Events have their place. But most kaizen should be smaller, faster, and embedded in daily work.

Two types of kaizen

• Daily kaizen: Small improvements executed within the team’s authority. An operator notices that the tool sequence during changeover wastes 30 seconds. She rearranges the tools on the shadow board. Tests it next changeover. Works. New standard. Done. This is kaizen. No event, no charter, no facilitator. PDSA cycle measured in hours.
• Project kaizen (event): Focused effort requiring dedicated team time and cross-functional resources. SMED study, cell redesign, flow improvement. These are 3-5 day events with a charter, a target, and a team pulled from regular duties. Project kaizen is necessary when the improvement crosses departmental boundaries or requires concentrated effort.

The ratio should be 80/20: 80% daily kaizen, 20% events. Most organizations invert this — they run occasional kaizen events and do no daily kaizen in between. The result: improvement happens in bursts with long gaps of stagnation.

Why daily kaizen matters more

A single daily kaizen saves 30 seconds on a changeover. Trivial. But 250 daily kaizens per year across a plant compound into thousands of hours saved, hundreds of quality improvements, and a workforce that thinks about improvement as part of the job — not as a special event that disrupts the routine.

PDCA (Plan-Do-Check-Act) is everywhere. PDSA (Plan-Do-Study-Act) is what Deming actually preferred. The difference is not semantic. It changes behavior.

Check vs. Study

• Check implies pass/fail. Did it work? Yes or no. If yes, great. If no, try something else. No learning required.
• Study implies learning. What happened? Why? What was unexpected? What would you change about the experiment? A failed experiment studied properly produces more value than a successful one checked off and forgotten.

The Study phase is where operational insight lives. When an experiment does not produce the expected result, the Study phase asks: was the hypothesis wrong, was the execution flawed, or did we learn something about the process we did not expect? Each answer leads to a different next step. Check gives you one answer (pass/fail). Study gives you understanding.

PDSA speed

A PDSA cycle should be as fast as the experiment allows. For a changeover improvement: Plan in the morning, Do at the next changeover, Study the result by end of shift, Act tomorrow. One day. Not one month. Not one quarter. The speed of PDSA cycles determines the speed of improvement.

A3 is a thinking discipline disguised as a form. The constraint of fitting everything on a single A3-sized page forces clarity. If you cannot explain the problem, root cause, and countermeasures on one page, you do not understand the problem well enough.

A3 as conversation

An A3 is not a form you fill out and submit. It is a conversation tool. The author brings the A3 to their manager. The manager reads it, asks questions, challenges the root cause, probes the countermeasures. “How do you know the root cause is the shadow board and not operator technique?” “Have you talked to the operators who do the fastest changeovers?” The A3 improves through this dialogue.

The manager’s job is not to approve the A3. The manager’s job is to develop the author’s thinking by asking questions that push the analysis deeper. This is mentoring through structured problem solving.

A3 is not for every problem

A3 is for problems that are important enough to require structured analysis and visible enough to need documentation. A tool falling off a shadow board does not need an A3. A chronic OEE problem affecting a hoshin project does. Use judgment. The overhead of A3 should be proportional to the significance of the problem.

The gemba is the place where value is created. The shop floor, the production line, the workstation. Kaizen happens in the gemba, not in a conference room. The people doing the work are part of the experiment, not subjects of it.

Rules of gemba experimentation

• Try it before you prove it. Do not spend three weeks building a business case for a $50 experiment. Try the countermeasure. If it works, scale it. If it does not, you learned something for $50 and one hour.
• Small scale first. One machine, one shift, one operator. Not the whole line. Small-scale experiments are fast, cheap, and reversible. If the experiment fails, you lose one changeover, not a week of production.
• Measure before and after. You cannot study a result you did not measure. Take baseline data before the experiment. Measure the same thing after. The comparison is the data. Opinions about whether it “feels better” are not data.
• Include the operators. They see things you do not. They know which step in the changeover is the bottleneck. They know why the material jams at that point. They know the workarounds that never made it into the standard work. Ask them. Listen. Include their ideas in the experiment design.
• Time-box the experiment. A kaizen experiment that runs for six weeks without a conclusion is a project, not an experiment. Set a time limit: try this for two changeovers and measure the result. If two changeovers are not enough data, run four. But decide within the week.

What rapid means

Rapid does not mean sloppy. It means biased toward action over analysis. Plan enough to have a hypothesis and a measurement. Do the experiment today. Study the result today or tomorrow. Act on the learning by the end of the week. If your PDSA cycle takes longer than a week for a shop floor improvement, you are over-planning or under-doing.

Kaizen does not exist in isolation. In the Campaign Framework, kaizen is the Execute phase — preceded by Shape (5S) and followed by Consolidate (audits, LSW). The execution flow connects the diagnostic tool (VSM) to the strategy system (hoshin).

How the flow works

1. VSM identifies the constraint. The current-state value stream map shows where flow breaks. Maybe it is a changeover bottleneck, a quality problem, a WIP accumulation. The VSM makes the constraint visible and quantified.
2. Kaizen burst targets the constraint. On the future-state map, a kaizen burst marks the improvement needed. “Reduce changeover from 28 to 12 minutes.” The burst has a specific target, not a vague aspiration.
3. A3 structures the problem. Before jumping to solutions, the A3 forces you to understand the current condition, analyze the root cause, and design countermeasures. This prevents the common failure of solving the wrong problem.
4. PDSA cycles execute the improvement. Try the countermeasure. Measure. Study. Decide. Iterate until the target is met or you learn enough to pivot. Multiple PDSA cycles may be needed for complex improvements.
5. Hoshin tracks the result. If the kaizen burst promoted to a hoshin project, the monthly savings track on the X-matrix. The improvement is connected to strategy, not floating independently.

Paradoxically, kaizen requires standard work. Not to bureaucratize improvement, but to ensure that each PDSA cycle produces usable knowledge.

What to standardize

• How you capture baseline data. What metrics? What sample size? How collected? If every kaizen measures something different, you cannot compare results or build institutional knowledge.
• How you document the experiment. Hypothesis, method, result, learning. Not a full report — a brief record. The A3 format works well for significant experiments. A notebook entry works for daily kaizen.
• How you update standard work. When a kaizen succeeds, the standard work procedure must change to reflect the new method. If the standard does not update, the old method creeps back. Someone must own the standard work update as part of every successful kaizen.
• How you share learning. The PDSA Study phase produces knowledge. Where does it go? If it stays in one person’s head, it is not organizational learning. Brief report-outs at tier meetings, A3 posting boards, shared repositories — the mechanism matters less than the discipline of sharing.

Kaizen events require a charter

For project kaizen (3-5 day events), a charter defines:

• Scope: What process? What boundaries? What is in scope and what is not?
• Target: Specific, measurable. “Reduce changeover time from 28 to 12 minutes.” Not “improve changeover.”
• Team: Who is participating? What roles? Who is the sponsor?
• Baseline data: Current performance. Without a baseline, you cannot measure improvement.
• Schedule: Day-by-day plan for the event. What happens each day?

Kaizen that does not connect to strategy is random improvement. It might make an area better, but “better” relative to what? If the hoshin objective is OEE > 85% and your kaizen improved inventory accuracy in the warehouse, you improved something that does not move the needle on the breakthrough objective.

Strategic vs. local kaizen

• Strategic kaizen connects to the X-matrix. It is a hoshin project with savings tracking, a charter, and monthly accountability. The improvement addresses a constraint identified by VSM that directly supports a strategic or annual objective. Examples: SMED to improve OEE, scrap reduction to hit yield targets, cell redesign to enable single-piece flow.
• Local kaizen operates within the MDI system. It does not require hoshin-level tracking. The team lead assigns it, the team executes it, and it is tracked on the newspaper. Examples: rearranging a workstation, adjusting a fixture, changing a material staging sequence. Important, but local.

Both types are necessary. Strategic kaizen drives breakthroughs. Local kaizen builds the improvement muscle that makes strategic kaizen possible. An organization that only does strategic kaizen events has an improvement program. An organization that does daily local kaizen has an improvement culture.

How to decide which type

• Does the improvement support a hoshin objective? → Strategic
• Does it require cross-functional resources or budget authority? → Strategic
• Can the team execute it within their authority? → Local
• Is the expected savings above the hoshin threshold? → Strategic

Most kaizens should be local. The threshold for hoshin promotion is typically $25K+ in expected annual savings, though this varies by organization. The X-matrix should have three to five projects per site, not thirty.