Insyt
All charts in the library

Flow, process and networks

Fishbone / Ishikawa cause-and-effect diagram

Organises possible causes of a problem into categories for structured investigation.

  • Diagram
  • Cause exploration
  • Problem structuring

What it answers

What possible causes should be investigated?

Shape
Spine with branching cause categories
What it shows
Candidate causes and gaps in investigation.

Example

The question

What could be causing late deliveries?

Possible causes of late deliveries (fishbone diagram)

Possible causes of late deliveries (fishbone diagram)The problem, late deliveries, sits at the head; six cause categories branch off the spine: People: new drivers, sick leave; Process: late picking, manual routing; Systems: address errors, slow scanners; Suppliers: late stock, wrong labels; Weather: snow, storms; Vehicles: breakdowns, too few vans.LatedeliveriesPeopleNew driversSick leaveProcessLate pickingManual routingSystemsAddress errorsSlow scannersSuppliersLate stockWrong labelsWeatherSnowStormsVehiclesBreakdownsToo few vans
Possible causes of late deliveries (fishbone diagram)The problem, late deliveries, sits at the head; six cause categories branch off: People: new drivers, sick leave; Process: late picking, manual routing; Systems: address errors, slow scanners; Suppliers: late stock, wrong labels; Weather: snow, storms; Vehicles: breakdowns, too few vans.Late deliveriesPeopleNew drivers, Sick leaveProcessLate picking, Manual routingSystemsAddress errors, Slow scannersSuppliersLate stock, Wrong labelsWeatherSnow, StormsVehiclesBreakdowns, Too few vans
The problem, late deliveries, sits at the head; six cause categories branch off the spine: People: new drivers, sick leave; Process: late picking, manual routing; Systems: address errors, slow scanners; Suppliers: late stock, wrong labels; Weather: snow, storms; Vehicles: breakdowns, too few vans.

Illustrative data. It belongs to no company.

When to use it

Use it when

Structuring a causal investigation or workshop.

Avoid it when

Treating brainstormed causes as proven evidence.

How to read it

Read the defined effect, then cause categories and subcauses.

Note

Possible causes remain hypotheses until supported by evidence.

The data you need

Layout
Cause-category table
Fields
Effect: text; CauseCategory: text; Cause: text
One row is
One candidate cause under a category for a defined effect. What is the grain?
Rules
Keep the effect specific. Separate possible causes from verified findings.
Optional
ParentCauseID for subcauses,Evidence,Owner.
Preparation
Validate the input grain and missing values. Keep the effect specific. Separate possible causes from verified findings.
Units
Relationships and positions have no numeric unit by default. Label any counts, durations, probabilities or weights separately.
Example of the input layout: Cause-category tableA long table: one row per observation, few columns. New data adds rows.EffectCauseCategoCauseAABBC
A long table: one row per observation, few columns. New data adds rows.

Worked example

EffectLate delivery; Process causeManual approval; Systems causeMissing alerts. Place both on branches and collect evidence before claiming either caused thedelay.

EffectCauseCategoryCause
Late deliveryProcessManual approval
Late deliverySystemsMissing alerts
All charts in the library