DMAIC|4 min read
5-Why analysis: when it works and when it doesn't
5-Why is Toyota's tool for drilling from a symptom to a root cause. Ask 'why?' five times. In theory it's simple. In practice, most teams stop at why #2 or blame a person at why #3. Both kill the analysis.
The right way
Good 5-Why chains follow the process, not the people. Example:
- 1. Why did the vial reject? → Seal pressure was out of spec.
- 2. Why was seal pressure out of spec? → Torque driver drifted during B-shift.
- 3. Why did the torque driver drift? → No hourly verification against master.
- 4. Why no hourly verification? → SOP missing the verification step.
- 5. Why is the SOP missing the step? → Last revision skipped a supplier change requirement.
Where it breaks
- Stopping at why #1 or #2 because the first cause 'feels' right
- Blaming an operator at why #3 ('operator didn't follow procedure') instead of asking why the procedure allows the error
- Jumping to solutions before hitting the true cause (premature countermeasure)
- Using a single 5-Why chain for a complex problem that actually has 3 root causes
When fishbone beats 5-Why
Use fishbone when the cause could be in any of the 6Ms. Use 5-Why when you already know the category and need to drill down. For a chronic, complex problem: fishbone first, then 5-Why on the top 2-3 candidates.
Confirm, don't assume
A 5-Why chain is a hypothesis chain. Why #5 is not a confirmed root cause until data from the process proves it. Never present a 5-Why in a tollgate without the evidence that validated the final why.