Ask Art
"What is mizen-boushi?"
What Is Mizen-Boushi?
Short answer: Mizen-boushi (未然防止) means prevention before occurrence — anticipating a failure and preventing it before it ever happens. Mizen means “not yet occurred”; boushi means “prevention.” It is the discipline of stopping a problem you have not had yet, and it lives upstream, in design and development.
What makes mizen-boushi different from reactive problem solving?
Tatsuhiko Yoshimura (吉村達彦), who spent 32 years at Toyota in reliability engineering and chassis development before moving to Kyushu University, laid out the progression clearly in his 2002 book Toyota-Style Pre-occurrence Prevention Method · GD³ (トヨタ式未然防止手法・GD³, JUSE Press):
- Problem solving (問題解決) — a problem has occurred; you investigate the cause and resolve it.
- Recurrence prevention (再発防止) — you prevent that same problem from happening again.
- Pre-occurrence prevention (未然防止) — you predict the occurrence of a problem that has not yet happened and prevent it before it occurs.
The first two are reactive — the problem already happened, and usually someone already paid for it. Mizen-boushi is the other direction: foresee the failure mode and design it out before the first occurrence, before a part is built, before a process is run. The whole point is in the word — mizen, not yet occurred.
Yoshimura noted that when he explained this in the United States, English had no word distinguishing pre-occurrence prevention from recurrence prevention. American engineers at Toyota advised him to use “mizen-boushi” as is, and Toyota adopted it as a global term.
Where does mizen-boushi live in the organization?
Because it is anticipatory, mizen-boushi occurs in design and development, not on the line. It is done in design review, where the process is still on paper and the cost of prevention is lowest. Changing a line on a drawing costs nothing. Changing a weld specification after ten thousand parts are in the field costs everything.
But “review the design for what could go wrong” is too vague to be useful — it invites boiling the ocean. Yoshimura’s critical insight was that you do not need to think of everything. Problems do not appear randomly across a design. They cluster at three specific types of discontinuity:
- Design changes — anywhere a designer deliberately departed from a proven design. A new geometry, a different material, a revised tolerance.
- Changed conditions — anything that changed on its own, even if the part’s design stayed the same. A different supplier, a new market environment, a neighboring component that was redesigned.
- Interfaces — the joints between parts, between supplier and customer, between one design department and another. Every boundary is a potential mismatch.
This is the focusing mechanism that makes mizen-boushi practical. Instead of asking “what could possibly go wrong with this entire design,” you ask “what changed, what changed around it, and where do things connect?” That question is answerable. It concentrates the design review on the places where problems actually originate — and keeps the team from wasting time reviewing dimensions that have not changed and have proven field history.
What are Toyota’s methods for mizen-boushi?
Yoshimura developed GD³ — Good Design, Good Discussion, Good Dissection — as a systematized approach to mizen-boushi.
Good Design starts from the premise that the basis of reliability is not changing things. If you do not change a proven design, reliability problems do not arise. The discipline is to minimize change points, make every change visible, and surface the discontinuities — between old design and new, between supplier and customer, between part and part — where the sprouts of problems hide.
Good Discussion is the structured design review focused on what changed and what could go wrong because of that change. Yoshimura found that conventional FMEA had become bureaucratic form-filling — engineers filled in boxes without actually discussing the risks. He reframed the FMEA worksheet as a tool for creative problem-finding through discussion, and combined it with a design review into a method called DRBFM — Design Review Based on Failure Mode. The detailed DRBFM implementation procedure was developed by Hirokazu Shimizu (清水浩和), a staff engineer at Toyota’s 5th Development Center.
Good Dissection is looking hard at the actual tested part — not the data sheet, the physical thing. After testing, engineers disassemble, observe, compare with untested parts, and ask what else might be wrong. Yoshimura formalized this as DRBTR — Design Review Based on Test Results. Former Toyota chairman Eiji Toyoda put it simply: engineers must look at things carefully, without preconceptions, and see what the object itself tells them.
Yoshimura systematized GD³ in his 2002 book after leaving Toyota for Kyushu University. The book’s foreword was written by Akihiko Saito, then Executive Vice President of Toyota, who called mizen-boushi “part of Toyota Motor Corporation’s DNA.”
Why is mizen-boushi difficult?
Yoshimura was honest about this. He described the reliability engineer as “a fortune-teller with no customers” — solving a problem in front of you earns gratitude, but preventing a problem no one has seen yet earns nothing, because people assume you took too much margin. His key insight after years of failures: the basis of pre-occurrence prevention lies in how one discovers failures. If recurrence prevention is interpolation — joining knowledge within a near, narrow region of past experience — then pre-occurrence prevention is extrapolation: joining distant knowledge across a wide region. The difference is small in principle and large in practice, which is why the structured methods exist.
How does mizen-boushi relate to poka-yoke and jidoka?
Mizen-boushi is the upstream end of prevention. On the line, poka-yoke guards against the inadvertent slips an operator can make in a process that already exists — but it cannot touch a flawed design or a wrong specification. Those are planning failures, and they have to be prevented earlier, which is exactly what mizen-boushi does. The two are not competitors; they are different layers of the same prevention picture, and mizen-boushi is the one furthest upstream.
See also: Jidoka — Part 1, Jidoka — Part 2, and Does Lean Forget Quality at Times?.