Knowledge · Strategy & Planning

Fishbone Diagram · Cause-and-Effect Diagram

Fishbone Diagram / Ishikawa Diagram

📂Strategy & Planning⏱2-min read📊Beginner🔗5 related

What is a fishbone diagram?

In one line
A root cause analysis tool that branches the causes of a single outcome (the effect) into a fishbone shape, so every candidate cause can be laid out and tested.
⚡ Summary in 3 lines
  • One outcome, with cause branches spread out like fish bones
  • Manufacturing: 4M/6M; services: people, process, systems, policy
  • It lays out the possibilities; data has to confirm the cause

The moment a fishbone goes up on the wall, the order of the meeting flips. Normally someone opens with the conclusion, “It’s just a lack of training,” and everyone else trades defenses and rebuttals from there. A fishbone pushes the conclusion back 30 minutes and has everyone lay out all the candidate causes first. What this tool really changes is that the branches, not the loudest voice, set the order of discussion. A team lead whose branch has only two or three sticky notes will go find more causes on their own before anyone points it out. It is this pressure, not the diagram itself, that produces results.

Four to six main branches is the limit. Manufacturing sites usually start with 6M and service organizations with 4P, but you can recast them in your own team’s terms, such as “people, process, systems, policy.” Forcing 6M onto office work leaves the Machine branch empty and makes the discussion awkward. Two levels of sub-branches is realistic, and any branch that needs to go deeper than that is better split off and handed to 5 Whys. It is easy to confuse with a logic tree: a logic tree demands a breakdown with no overlap, while a fishbone is a divergent tool that lets the same cause sit on two branches. It is normal for 30 to 50 candidates to come out at once, and the number actually worth testing usually narrows to 5 to 7.

The most common mistake is to draw everything and then pick the “real cause” by show of hands. The vote usually matches whichever branch was loudest in the room, and countermeasures built on that cause let the same problem come back three months later. Hold one line firmly: data narrows the candidates; the diagram only produces what needs narrowing. Stating the effect vaguely, as in “weak sales,” leads to the same failure. When the problem at the head is too big, the branches simply copy the company org chart, and the workshop ends two hours later with “let’s all do better.” Put up an effect narrowed with numbers to when it got worse and by how much, and the sub-branches go down to a level you can verify.

Courses That Cover This Term · Finance & Strategy
Recommended courses · Recommended instructors · Quick quote
🧮 Mini quote · Based on instructor fees
About KRW 1.2 million (≈ US$900)2-hour talk · up to 20 people · Korean
Get a quote for these requirements →
💡 In plain English
It resembles a medical exam: rather than guessing at the cause of an illness, the doctor reviews every item on the intake form and then decides which tests to run.
Why it matters in practice
In a meeting culture where the loudest voice wins the cause debate, it is a practical facilitation tool that gets every department looking at the same picture and laying out causes together.
Origin and sources

Kaoru Ishikawa (石川馨), a University of Tokyo professor and quality-control authority, devised it in the 1960s while advising Kawasaki Steel, which is why it is also called the “Ishikawa diagram.” Along with Deming and Juran, he led Japan’s quality revolution, and the diagram became one of the seven basic QC tools.

Source · Kaoru Ishikawa, “Guide to Quality Control” · View source
An example

Consider a hypothetical case. At Company A, a food manufacturer with 320 employees and annual revenue of about KRW 140 (≈ US$1) billion, finished-product foreign-matter claims rose from 4 a month to 13 a month in three months. Production, quality, and maintenance kept blaming one another, and two response meetings ended with nothing. The QA manager then gathered 11 people, including the floor supervisors, for a 2-hour fishbone workshop.

  1. When the QA manager wrote “Packaging foreign-matter claims: 4 to 13 a month, Aug–Oct” at the head of the diagram, the production manager said, “Then we only need to look for what changed after August.”
  2. The team set five main branches (people, equipment, materials, methods, measurement), and each member wrote causes on sticky notes. In 20 minutes, 41 candidates filled the wall.
  3. The equipment branch surfaced “sealing temperature variation on Line 3,” the people branch “7 new hires added in August,” and the materials branch “a change of packaging film supplier.”
  4. When the team lead said, “Let’s keep only what actually changed since August,” the list shrank from 41 items to 6. The team then pulled the original claim records for the past three months and checked how many claims each item accounted for.
  5. The change of film supplier overlapped with 9 of the 13 claims. Applying 5 Whys to that branch showed that adhesion strength was missing from the incoming inspection items.
  6. After the team added inspection items and reset the temperature control standard for Line 3, foreign-matter claims fell from 13 a month to 3 within two months.
Related terms
Training linked to this term
📚Problem Solving & Crisis ManagementNanum Consulting corporate training · On-site or onlineView course →
Need in-house training on this topic?
Get a curriculum and quote tailored to your needs right away.
Ask About Training
Last updated 13 Sep 2026 · Reviewed by John Kim

FAQ

No. Choose categories that fit your organization’s language: 4M/6M for manufacturing, or people, process, systems, and policy for services. What matters is that the branches don’t overlap and cover the whole picture.
No. A fishbone lays out every candidate cause; confirming the real one is the job of later steps such as data validation and 5 Whys.
You can, but the effect is halved. Each department knows different causes, and the information gaps that surface while drawing together are a large part of what makes this tool useful.
Screening by two criteria, “what changed recently” and “what we can control,” usually narrows the candidates to 3–5.