A fishbone diagram organises possible causes around a clearly defined effect. Also called a cause-and-effect or Ishikawa diagram, it helps a team widen its enquiry. The branches contain hypotheses until investigation supports them; the diagram itself does not establish which cause produced the problem.
The Corrective Action and Continual Improvement in Quality Management places this topic in its wider QMS context. Category labels can be adapted to the process. Manufacturing categories are not compulsory for an office or service investigation.
Define the decision and its boundary
Before designing controls for fishbone diagrams, understand the current practice. Compare the written arrangement with a real case and ask where people need informal help. Do not assume that a difference automatically indicates misconduct or a poor procedure. Determine whether the practice protects the intended result, creates an uncontrolled dependency or reveals that the official description needs revision.
Invite people with different perspectives before finalising categories. A diagram developed only by managers may overlook what happens during real handovers, while a team from one shift may miss conditions elsewhere. Capture disagreement as a hypothesis to investigate. Retain a readable version showing which branches were tested and what was learned. A final diagram with every idea left unqualified does not show the outcome of the enquiry.
Put the method into practice
Organise fishbone diagrams around decisions rather than signatures. Explain what each participant must establish before handing over the work, including unresolved limitations. A second approval is useful only when it adds relevant expertise, authority or challenge. Remove duplication carefully, checking that the revised arrangement still makes responsibility clear and leaves the receiving person able to act.
-
Write one specific effect at the head of the diagram.
-
Gather possible causes under categories suited to the work.
-
Prioritise hypotheses for evidence collection and mark investigation outcomes.
Worked example
A fictional support team investigates reopened cases. It groups possibilities under information, tools, skills, workload and handovers. One branch suggests that closure messages omit a required customer confirmation.
Review shows several reopenings concern incomplete resolutions, while others involve new requests. The team separates these populations before testing whether its closure process needs revision.
The worked example separates an observation, an interpretation and a response in fishbone diagrams. Preserve that distinction when discussing a real event. People can agree that an output was unsatisfactory while disagreeing about why it happened. Obtain evidence that addresses the disagreement, and avoid reporting an appealing explanation as an established fact before the relevant conditions have been examined.
Useful evidence and records
Use a proportionate record of fishbone diagrams that distinguishes facts, judgement and outstanding work. The same information should not be entered repeatedly merely to satisfy different report layouts. If summaries are needed, preserve traceability to the underlying case. Decide who updates the status and how corrections remain understandable, particularly when several functions participate in the decision.
| Record element | What it should establish |
|---|---|
| Effect | Specific problem and population |
| Branch | Potential mechanism and category |
| Evidence | Support or contradiction found |
| Disposition | Investigate, reject or retain uncertainty |
Check the result and avoid a common mistake
A crowded diagram can create an illusion of thoroughness. Unsupported branches must not be presented as proven causes, and voting does not establish causality.
Select a small number of credible branches and define the observations needed to distinguish them. Update the diagram as evidence changes the explanation.
Include challenge in the review of fishbone diagrams. Ask what evidence would lead the team to revise its conclusion and look for that evidence deliberately. A review designed only to confirm success is unlikely to reveal an important limitation. Record a bounded conclusion where appropriate, explaining which conditions were tested and what follow-up is needed for the remaining uncertainty.
Connect the next decision
Make the next review of fishbone diagrams purposeful. Identify the event or information that would justify reconsidering the current approach and assign someone to recognise it. Routine review dates can support this arrangement, but they should not prevent earlier action when conditions change. Close the current discussion with a clear decision and a practical route for any remaining enquiry.
Use CAPA Explained in Quality Management when the work moves to the adjacent decision. The related guide to root cause analysis explains the wider process dependency. The handover should identify what has been accepted, what remains open and who will review the next evidence. Keep the original decision traceable when later information changes the conclusion.
Can a completed diagram close the investigation?
No. Completion shows that possible causes were organised, not that any branch was verified. The investigation record should identify which explanations were tested, what observations supported or contradicted them and why the chosen response follows from the evidence. Preserve unresolved limitations for later review.