The template
| Category | Prompts | Causes raised |
|---|---|---|
| Machine (equipment) | Settings, wear, maintenance, capability, changeover | |
| Method (process) | Standard exists? Followed? Sequence, timing, handoffs | |
| Material | Supplier, batch variation, storage, specification | |
| Measurement | Gauge accuracy, sampling, how the defect is detected | |
| People (manpower) | Training, sign-off, fatigue, staffing, communication | |
| Environment (mother nature) | Temperature, humidity, light, noise, housekeeping |
Problem statement (the fish head): one measurable effect, for example "Label jams: 14 per week on Line 3, 11 after changeover."
How do you run a fishbone session?
- Write the problem statement with numbers. If you cannot, go back and measure.
- Draw the spine and the six bones on a whiteboard or large sheet.
- Brainstorm causes under each bone with the people who do the work. Every cause is written; none is argued yet.
- Ask why of each cause until it reaches something you could change; add the answers as smaller bones.
- Circle the three to five causes the team thinks most likely.
- Go and check them: data, observation, a trial. The diagram ends when causes are confirmed, not when it is full.
What are the common mistakes?
- A vague head ("quality problems") that no cause can be tied to.
- Doing it in a meeting room without operators.
- Stopping at "operator error" or "machine fault" without asking why.
- Treating the finished diagram as the analysis. It is the list of suspects; the analysis is confirming which one did it.
Fishbone, 5 whys and Pareto
A Pareto chart chooses which problem to attack. The fishbone organises the possible causes of that problem. The 5 whys drives each likely cause down to a root. Used in that order they are the analysis section of an A3.
Fishbone diagrams and daily management
A fishbone is the first tool reached for when a measure on the tier board turns red for the same reason twice. The session takes thirty minutes at the board with the team, the confirmed causes become actions with owners, and the result is reviewed at the next huddle.
Step by step
- State the problem with numbers. One measurable effect in the fish head.
- Draw the six bones. Machine, method, material, measurement, people, environment.
- Brainstorm causes. With the people who do the work; write everything, judge nothing.
- Ask why. Drive each cause toward something changeable; add the answers as smaller bones.
- Select and confirm. Circle three to five likely causes and test them with data or observation.
A worked example
A dairy filling line had seal failures on 3 per cent of cartons. The fishbone session with two operators and the maintenance fitter produced 31 causes; the team circled five. Checking the data killed three within a day. The remaining two, seal-bar temperature drifting after the mid-shift clean and a new carton supplier batch, were confirmed by a trial on one shift. Both became actions: a temperature check added to the post-clean checklist and a supplier batch hold. Seal failures dropped below 0.5 per cent in three weeks.