Root cause analysis: 5 whys and Ishikawa done right, with a real example and the most common mistakes
The 5 whys and the Ishikawa diagram are the two most taught and most misused root-cause tools: they get filled in ten minutes, land on "lack of training" or "human error" and the problem returns next month. This guide shows how to do them right, with a real late-delivery example and the mistakes that make them useless.
In this article
What a root cause is and is not
The root cause is the condition that, if removed, prevents the problem from recurring. It is not the first explanation that appears ("the driver was late"), nor the most comfortable ("lack of training"), nor the one that blames a person. A simple test: if the "cause" you found can be fixed with a reprimand, it is not the root.
Before asking why: define the problem with data
Before asking why, define what happens, where, when and how much. "Deliveries arrive late" is not a defined problem. "18% of north-route deliveries arrived after the committed time in September, versus 4% on other routes" is. Without data, the 5 whys get answered with opinions.
The 5 whys, step by step
You ask "why?" in a chain until you reach a cause you can act on. With the example:
- Why do north-route deliveries arrive late? Because trucks leave after 9:00.
- Why do they leave late? Because loading finishes after 8:30.
- Why does loading finish late? Because that route's orders are released last.
- Why are they released last? Because the system sorts orders by folio and the north route enters its orders later.
- Why does it enter them later? Because its customers send orders until 7:00 and the cutoff rule is not defined.
Root cause: there is no order cutoff time per route. Action: define the cutoff and sort loading by route and departure time. Notice that "train the drivers" appears nowhere.
The Ishikawa diagram, when and how
The 5 whys follow a single line; when the problem has several possible causes, the Ishikawa (fishbone) diagram organizes them by category before going deep. The classic categories are method, machine, manpower, material, measurement and environment, but it pays to adapt them to the operation (in services: process, people, system, information, customer, environment).
How to do it right: (1) the data-defined problem at the head of the fish; (2) as a team, brainstorm possible causes by category, without judging; (3) mark the three or four most likely causes; (4) verify each one with data before acting; (5) apply 5 whys to the verified ones. The diagram does not give the cause: it gives the list of hypotheses.
Verify before acting
It is the step almost everyone skips. A cause is a hypothesis until data confirms it: if you believe the order cutoff is the cause, check what time the orders of late deliveries were entered versus on-time ones. If there is no difference, the cause is elsewhere. A Pareto of verified causes says where to start.
The five most common mistakes
| Mistake | What it looks like | How to fix it |
|---|---|---|
| Problem without data | "Deliveries arrive late" | What, where, when, how much, compared to what |
| Stopping at "human error" | "The driver made a mistake" | Ask why the process allowed the error |
| Stopping at "lack of training" | Comfortable cause that is never verified | Ask for evidence: did someone who knew how still fail? |
| Doing it only at the desk | Causes invented far from the process | Go to the floor, talk to operators, look at the data |
| Not verifying the cause | Action that does not move the indicator | Check against data before the action plan |
From analysis to action plan
A verified root cause becomes an action with an owner, a date and an indicator that will say whether the problem disappeared (in the example: percentage of late north-route deliveries in November). A simplified one-page A3 holds it all and is the deliverable of the critical thinking and problem solving course; for plants that want the full improvement system, the lean manufacturing course complements it.
Want to bring this to your team?
Each team works a real problem from its operation start to finish: data-based definition, verified root cause and an A3 with an action plan. Message us on WhatsApp, fill out the contact form or open this site's live chat: we send a proposal with three scope options and a closed price in under 24 business hours, with the DC-3 certificate and Geedle Bot included. The full program is at the critical thinking and problem solving course and the catalog at training courses for companies.
