The fishbone diagram: find all the possible causes before you pick one

Run this method in IdeaHone
Creates a space with the entities and scoring criteria of this method, ready for your next session.
When something goes wrong, teams usually agree on a cause within minutes. The first plausible explanation wins, someone fixes it, and a month later the problem is back. The fishbone diagram slows that moment down. Before anyone picks a cause, the team collects every possible one, sorted into categories, so the obvious answer has to compete with the others. Kaoru Ishikawa made the diagram popular in Japanese quality circles in the 1960s, which is why it is also called the Ishikawa diagram. The shape gives it the other name: the problem is the head, the categories are the bones.
How the diagram works
The problem goes in the head on the right, written as an observable fact: "The refill station is used by fewer than 5 customers a day", not "the refill station failed". A horizontal spine runs to the left. From the spine, bones branch off, one per category of cause. Possible causes are written on each bone, and causes of causes branch off those.
The categories are there to make people look in places they would otherwise skip. The original manufacturing set was the 6 Ms: man, machine, method, material, measurement and mother nature. For product and service teams, a version like this works better:
| Category | Ask about |
|---|---|
| People | Skills, roles, workload, communication |
| Process | Steps, rules, hand-overs, timing |
| Technology | Systems, machines, software |
| Information | Inputs, data quality, what people know or are told |
| Measurement | Whether you measure the right thing, the right way |
| Customers and market | Needs, habits, context, competitors |
Use the categories that fit your problem. Four to six is enough; more makes the diagram hard to read.
Running a fishbone session in 45 minutes
- Write the problem as a fact. What happens, where, since when and how much. If the team cannot agree on the problem, stop here; the rest will not help.
- Draw the bones. Pick four to six categories and write them on the bones before anyone names a cause.
- Write causes alone for ten minutes. Everyone writes possible causes, one per note, and places them on the bone they belong to. Possible, not proven: this is the moment for breadth.
- Read out and merge. Go bone by bone. Merge duplicates and ask "what causes that?" once for each cause, adding the answer as a smaller branch.
- Mark the likeliest causes. Three dots per person, for the causes they think are most likely. Not the ones easiest to fix.
- Check the top causes against facts. For each top cause, name what you would see if it were true, and go and look. "Bottles left at home" predicts that people who bring a bottle once come back; the till data can show whether they do.
- Dig into the confirmed cause. Take the cause that holds up into the 5 whys until you reach something the team can change, and give it an action with an owner.
Keep the board up for a week after the session. Causes people dismissed in the room often turn out to be right once someone looks at the data, and the board shows the team what it already considered.
Fishbone or 5 whys?
They work well together. The fishbone is wide: it collects many possible causes across categories, so the team does not lock onto the first one. The 5 whys are deep: they follow one cause down to a root the team can change. Start wide when the cause is unclear or disputed, then go deep on the cause that survives the check. When the cause is already obvious and agreed, the 5 whys alone are enough.
Where fishbone sessions go wrong
- A vague problem in the head. "Low adoption" invites every opinion in the building. A specific, measured problem keeps the causes specific too.
- Blame on the people bone. "Staff do not care" ends the conversation. Ask what makes the job hard instead: training, time, tools.
- Voting instead of checking. Dots show what the room believes. The check against facts shows what is true. Skip it and you fix the most popular cause, not the real one.
- A diagram and no action. The session ends with a decision about which cause to work on and who does what, or it was a nice drawing.
Run the workshop. IdeaHone keeps the record.
While you facilitate, IdeaHone turns what is said into ideas, questions and decisions on a shared board, with who said it. Nobody writes the minutes afterwards.
What is a fishbone diagram?
A cause and effect diagram that places a problem in the head of a fish and its possible causes on bones sorted by category. It helps teams collect causes widely before choosing which to investigate. It is also called an Ishikawa diagram, after Kaoru Ishikawa.
What are the 6 Ms?
The classic categories from manufacturing: man (people), machine, method, material, measurement and mother nature (environment). Service and product teams often adapt them, for example to people, process, technology, information, measurement and customers.
When should I use a fishbone diagram instead of the 5 whys?
When the cause is unclear, disputed, or likely to have several parts. The fishbone collects possible causes across categories; the 5 whys then dig into the one that holds up against the facts.
How long does a fishbone session take?
About 45 minutes for the diagram, plus the time to check the top causes against data afterwards. That check often takes a few days and is the part that makes the session worth it.