Fishbone Diagram: Map Cause Families Before You Pick a Story
The Fishbone Diagram (Ishikawa, cause-and-effect) helps a team map a named problem to categories of candidate causes, then choose a branch to test. It is a wide map, not a single why-chain and not proof that a cause is true.
Framework Card
- Name:
- Fishbone Diagram
- Goal:
- Map a named problem to categories of candidate causes so a team can see more than one story before they pick a branch to test.
- Flow:
- Problem (effect) → Cause categories → Candidate causes per category → Branches to test
- Best For:
- Exploring a messy problem with several possible causes; Running a team diagnosis without jumping to one favorite story; Preparing a branch for a deeper why-chain
Why it matters
Repeating problems attract a favorite story: “it is training,” “it is the vendor,” “it is the machine.” The story may be early, and it may be incomplete.
When several causes can operate at once, a linear argument skips whole families of causes. The meeting then fights personalities instead of filling a map.
Fishbone is a way to slow that fight. You put the effect at the head, you pick bones (categories), and you hang specific candidate causes on each bone. The useful output is not a pretty skeleton. It is a short list of branches that deserve evidence.
What it is
Kaoru Ishikawa popularized the diagram in quality work. The head is the problem. The bones are cause categories. A common manufacturing set is Equipment, Process, People, Materials, Environment, and Measurement (often called 6M). Other work uses different bones. The method is the categorized map, not a mandatory six labels.
Fishbone organizes hypotheses. 5 Whys drills one line. Keep them distinct: use Fishbone when the problem is wide enough that several cause families may be in play; use 5 Whys when you are ready to push one branch deeper.
How it works
1. Name the problem
Write a specific effect (“orders ship with the wrong SKU,” not “warehouse is sloppy”). That sentence is the head.
2. Choose categories
Pick bones that fit the work. Manufacturing 6M is a default, not a law. Service teams might use people, process, policy, systems, vendors.
3. Populate causes
Under each bone, list candidate causes that could produce this effect. Stay specific (“no barcode scan at pack,” not “people”).
4. Mark branches to test
Mark which causes have evidence, which are guesses, and which deserve a deeper probe (often 5 Whys on one bone). Without that cut, you have a poster.
How it compares
When another lens fits better, or when you need a complementary view, these frameworks do different jobs.
| Framework | What it helps you see | How it differs from Fishbone |
|---|---|---|
| 5 Whys | One cause chain by asking why | Deep on one line. Fishbone is a categorized map. |
| FMEA | Failure modes scored for risk | Prospective risk ranking. Fishbone maps causes of a named effect. |
| 80/20 Rule | Which counted items dominate a result | Ranking with evidence. Fishbone lists hypotheses. |
| PDCA Model | Plan, do, check, act | Cycle after you have a change to try. Fishbone is the map first. |
The Fishbone Diagram is the lens for categorized cause hypotheses. Other methods help when the question is one chain, scored failure modes, concentration, or an improvement loop.
When to Use This Framework
- Complex cause exploration. Symptoms repeat and more than one family of causes is plausible.
- Team-based diagnosis. People will otherwise argue one favorite explanation.
- Root cause mapping. You need categories before you pick a branch to test.
Example
A concrete example makes the structure easier to reuse when you are under uncertainty.
Example: Late outbound orders
A fulfillment team misses the same-day cutoff three days a week. Favorite story: “staffing.”
They put missed same-day cutoff at the head. Bones (illustrative, not a universal 6M law):
- People: overtime fatigue on Thursday; two packers untrained on the new scanner
- Process: cutoff is 3 p.m. but pick wave starts at 2:30
- Equipment: one scanner drops Bluetooth twice an hour
- Materials: labels print with yesterday’s batch ID after a reprint
- Environment: packing lane blocked by inbound pallets at 2 p.m.
- Measurement: “on time” is counted at carrier scan, not at pack complete
Implication: Staffing may still matter. The map also shows process timing, a device, a label bug, and a blocked lane. The next move is to test the highest-evidence bones, not to stop at “people.” This is not a counted Pareto and not a finished 5 Whys.
Takeaway
What the Fishbone Diagram can help with
- Making several cause families visible at once
- Giving a team a shared map instead of one favorite story
- Preparing a specific branch for evidence or 5 Whys
- Reducing blame-only explanations that skip process and measurement
What the Fishbone Diagram cannot replace
- 5 Whys. One causal chain. Fishbone is wide. 5 Whys is deep on a chosen line.
- FMEA. Failure modes with severity, occurrence, and detection (or another risk ranking). Fishbone maps causes of an effect you already see.
- 80/20. Ranking causes you can already count. Fishbone hypothesizes; it does not rank by frequency unless you add counts.
- PDCA. A cycle for trying a change. Fishbone diagnoses candidates; it does not run the experiment.
- Proof. A full fish without checks is still a set of guesses.
Honest scope: Fishbone structures a cause map. It should not be sold as the root cause by itself.
Frequently asked questions
Fishbone maps many possible causes across categories. 5 Whys pushes one line deeper. Use Fishbone when several families may be in play. Use 5 Whys when you are ready to test one branch. They often work together: map, then drill.
A specific problem statement, named categories, specific candidate causes, and marked branches to test or evidence. If every bone says “communication” or “training,” the map is still slogans.
No. 6M is a common teaching set. Choose bones that fit the work. Empty mandatory categories help no one.
No. It proves you considered more than one family. Causes still need evidence.
No. FMEA lists how something could fail and ranks risk. Fishbone explains an effect you already have by grouping candidate causes.