Decision-makingAlso called Five whys, why-why analysis
5 Whys
Ask why the problem happened, then ask why again of each answer, until you reach a cause worth fixing.
In one look
- 1Start with one problem.
- 2Ask why of each answer.
- 3Fix the cause you reach.
It comes from Toyota. The habit is credited to Sakichi Toyoda, and Taiichi Ohno made it standard practice in the Toyota Production System. His example: a machine stops because a fuse blew, because a bearing ran dry, because the pump failed, because its shaft wore out, because there was no strainer to keep metal scrap out. Fit the strainer and it stays fixed.
The problem it solves
"Human error" has closed this ticket three times.
The review finds someone who made a mistake, sends a reminder and closes the ticket. It happens again in March. Asking why five times sounds easy. It's harder in a meeting that already has an answer it likes, and harder still to keep the questions on the process and off the nearest person.
Try it yourself
No sign-up, a few minutes. A routine command just took part of a cloud service offline. You follow the whys and pick the root cause.
The same incident, five whys down
A proposal went out with the wrong pricing.
- The ticket saysCause: human error. An old template was used.
becomes The old template is the top search result on the shared drive.
- The ticket saysAction: reminder sent to check templates.
becomes Nobody is responsible for removing old versions.
- The ticket saysStatus: closed.
becomes The team outgrew "everyone just knows." Templates get an owner.
Good times to run it
- Reviews of a failure that keeps coming back
- A client deliverable that goes wrong the same way each time
- Support tickets that cluster around one step
- Before anyone proposes new software as the fix
When to skip it
- Big, tangled failures. They rarely have one root cause, and a single chain hides the rest.
- When only one team is in the meeting. A different team would give you a different chain.
- Pointed at a person. Then it's blame.
Teams we've run sessions with:
More frameworks
- Decision-makingPre-MortemBefore the work starts, the team imagines the project has already failed and explains why.
- Structuring thinkingIssue TreesBreak one big question into smaller questions that don't overlap, until each is small enough to answer.
- Structuring thinkingFirst Principles ThinkingBreak a problem down to the facts you can check, then reason up from those instead of from what's usually done.