Fishbone and FMEA point in opposite directions in time. A fishbone diagram looks backward: a problem happened, and you map every cause that might explain it. An FMEA looks forward: nothing has failed yet, and you score every way it could, so you can prevent the worst ones before launch. The fastest way to choose: has the problem already happened? If yes, you are investigating — reach for the fishbone. If you are trying to stop failures that have not happened yet, especially with safety or regulatory stakes, you need FMEA.
Side by side
| Fishbone (Ishikawa) | FMEA | |
|---|---|---|
| What it is | Qualitative cause-mapping diagram | Structured risk-assessment worksheet |
| Direction | Reactive — a problem occurred | Proactive — failures that have not happened |
| Question it answers | "Why is this happening?" | "What could fail, how bad, how likely, how detectable?" |
| Scoring | None — causes are unranked | RPN = Severity × Occurrence × Detection |
| Output | Branching 6M cause map | Ranked worksheet + actions + residual RPN |
| Driver | An incident or a curiosity | Risk prevention, often regulatory |
| Team | 4-8 person brainstorm | Largest cross-functional group of any RCA tool |
| Lifespan | One-time artifact | Living document, updated over product life |
| Time | 1-2 hours | 2-20+ hours, recurring |
The key insight: same raw material, opposite jobs
Both tools enumerate causes — that is why they look similar at a glance. The difference is what they do next. A fishbone stops at the map: its purpose is to understand a real problem well enough to act. An FMEA keeps going: it scores each potential failure by how severe it is, how often it is likely to occur, and how likely you are to catch it before it reaches the customer — multiplying the three into a Risk Priority Number that ranks where prevention effort belongs.
That difference is also why they pair so naturally. A fishbone session is one of the cleanest ways to fill an FMEA's "potential cause" column without missing whole categories — the 6M branches force breadth before you start scoring. Run the fishbone as the front-end brainstorm; let the FMEA be the back-end risk register that scores, prioritises and tracks the actions. The fishbone asks what could go wrong; the FMEA decides which of those is worth preventing first.
Choose Fishbone when…
- A problem already happened and you need to understand it fast. Fishbone is built for the reactive case — FMEA's scoring machinery is overkill for a single known incident.
- You want a shared visual for a brainstorm. The diagram aligns a cross-functional team on one picture in an hour.
- Stakes are modest and nobody will audit the output. No RPN, no detection column, no living document — just a clear map of suspects.
- You will drill with 5 Whys next. A fishbone hands 5 Whys the candidate branches to pursue.
Choose FMEA when…
- You are designing a product or process and want to catch failures before they ship. FMEA is a prevention tool first — its whole value is acting on failures that have not happened yet.
- You are in a regulated industry. IATF 16949, AS9100 and FDA guidance expect a documented, scored FMEA; a fishbone will not close the requirement.
- You need to rank risks and justify a prevention budget. RPN turns "this worries me" into a defensible priority order.
- The cost of a missed failure is high — safety, recall, or major downtime. The Detection dimension exists precisely for these stakes.
The 30-second decision rule
Has the problem already happened? → Fishbone (then 5 Whys) to understand it. Trying to stop failures that have not happened yet — especially with safety or regulatory stakes? → FMEA, to score and prevent them. Designing something new points to FMEA; investigating something broken points to fishbone. When both apply, run the fishbone first and feed its branches into the FMEA's cause column.
Map causes with the free Fishbone tool →Two mistakes that waste days
Mistake 1: FMEA for a problem that already happened. Pointing the full scoring apparatus at one specific incident is slow and disproportionate — a fishbone plus a 5 Whys drill reaches the root cause in an afternoon. (The lesson learned should feed back into the relevant FMEA afterward, but that is the follow-up, not the investigation.)
Mistake 2: fishbone for high-stakes proactive risk. A brainstorm diagram has no scoring and no detection dimension, so it cannot tell you which potential failure deserves the prevention budget — and it will not satisfy an auditor. When failures have not happened yet and the stakes are real, an unscored cause map is not enough; you need the RPN.
Frequently asked questions
Can fishbone and FMEA be used together?
Yes — a fishbone is a common way to brainstorm the potential causes that populate an FMEA. Fishbone is the ideation; FMEA is the structured, scored risk register.
What is the main difference?
Direction and scoring. Fishbone is reactive and unscored (why did this happen); FMEA is proactive and scored by RPN (what could fail and how risky).
Is a fishbone part of FMEA?
Not formally, but it is frequently used to generate the failure-mode and cause ideas before they are entered and scored in the FMEA worksheet.
Which is better for regulated industries?
FMEA — it is required by standards like IATF 16949, AS9100 and FDA guidance. A fishbone alone will not satisfy those requirements.
When is a fishbone enough?
For reactive, lower-stakes investigation where you just need to understand causes quickly — no scoring, no audit trail, no regulatory requirement.
Related resources
- Fishbone diagram: the complete Ishikawa guide — 6M framework with a full worked example
- FMEA: the complete guide — RPN formula, rating scales, DFMEA vs PFMEA, regulatory context
- 5 Whys vs Fishbone — depth vs breadth on the reactive side
- Fishbone vs Fault Tree — qualitative brainstorm vs quantitative probability tree
- All five RCA tools compared — the full decision guide