Quality in Practice — Lesson 2: 5 Whys and Ishikawa: finding the true cause, not the culprit

Quality in Practice · 5 min read · 25.08.2026

The second lesson of the Quality in Practice series. In the first, we showed that quality is created in the process and final inspection only reveals what the process got wrong. When a defect appears anyway, you face a question: look for "who" or "why". In this lesson, we will walk through the 5 Whys method and the Ishikawa diagram: when to use which tool, how to lead the analysis with a team right on the shop floor, and where the most common mistakes are. The exercise at the end takes no more than half an hour.

The culprit ends the search exactly where it should have begun

A scene our consultants see on audits over and over: the foreman identifies a defect, assigns it to an operator, reprimands him and closes the nonconformity as resolved. Three weeks later, the defect is back. At an engineering company with 240 employees, we analyzed 20 repeat complaints; in 16, the original "cause" was recorded as human error. Not one of those analyses asked why the process allowed the error to occur and pass through.

The problem with hunting for a culprit is not moral, it is practical: whoever is punished for reporting stops reporting nonconformities. At one welding operation, after we introduced root cause analysis instead of reprimands, reported nonconformities rose from 11 to 34 in three months — and complaints fell by 40 %. The defective parts had been there before, too. Now somebody saw them at a time when something could still be done about them.

Human error is not a cause. It is where the search for the cause begins.

5 Whys: a simple technique, a hard discipline

The method is trivial: for every answer, ask "why" until you reach a cause you can change. The number five is just a guide — real chains have 3 to 7 links. Two rules are essential: every answer must be verified by a fact or a measurement at the place of work, not an opinion in a meeting, and the last "why" must point to the process, not to a person's character.

An example from a screwdriving line, where inspection found a torque of 18 Nm instead of the prescribed 25:

1. Why was the torque 18 Nm? The operator used a substitute wrench without adjustable torque.
2. Why did he use it? The prescribed preset torque wrench was broken.
3. Why was it not replaced? There is only one unit in the department and the repair took longer than the planned downtime.
4. Why is there only one? The spare is ordered through central purchasing, the lead time is 6 weeks, and nobody wants it on their plate.

The first answer sounds like carelessness; the fourth points to tool inventory, the ordering process, and responsibility. The solution: two preset torque wrenches for the department, a calibration check in the shift start-up checklist, and a duty to report broken tools within 24 hours. Torque complaints fell from 9 to 0 within a quarter. Had we stayed at "the operator used the wrong wrench", the defect would have returned with the operator on the next shift.

The most common mistakes with 5 Whys from our practice: the chain stops at a person, the answers are unverified hypotheses, the analysis is written by an engineer alone at a desk instead of by a team at the machine. If the chain ends with "not enough people" or "we have no money", ask what specifically led to this defect on this day. Such an answer is an escape route, not a cause.

Ishikawa: when a problem has more than one cause

5 Whys is linear — it assumes a single chain of causes and effects. In real manufacturing, most defects arise from a combination of circumstances. That is what the Ishikawa diagram, the fishbone, is for. A team of 3 to 5 people brainstorms possible causes in six categories: machines, methods, material, measurement, people, environment. Write down every hypothesis and do not pick any of them apart during the brainstorming.

An example: scratches on rolled bars. A team of five spent 25 minutes at the line and came up with 14 hypotheses. In two hours of measurements, they ruled out 9 and confirmed 3: a worn roll, changed emulsion viscosity after a new delivery, and a feed rate outside the standard after interrupted production. Each confirmed branch got its own 5 Whys. Scratches fell from 3.2 % to 0.4 % of production within two months.

We recommend this combination as the standard: Ishikawa for an overview of the possibilities, 5 Whys to dig to the root of each confirmed branch. Do not throw away the ruled-out hypotheses — write them down. With the next similar defect, you save an hour of speculation.

This week's exercise

Pick one recurring nonconformity from the past month — one that has come back at least twice. Take a sheet of paper to the place where it originates and bring the operator who does the work. State the goal up front: we are not looking for a culprit, we are looking for the reason. Ask "why" and write down the chain. Verify every answer with a fact: check the machine log, take a measurement, review the documentation. Keep going until the last answer is something you can change within a month. Then compare your chain with the original entry in the nonconformity report. Usually you will find that the original analysis went only one or two steps deep. It is precisely in this gap that repeat complaints live.

You can handle root cause analysis on a single line in one morning. If you want to build teams that do this across all your plants, take a look at our e-learning offer — series like this one are complemented there with workshops and templates. In the next lesson, we will move on to the 8D report: how to turn a complaint into a managed project, not a form filled in the day before the audit.

Quiz — test your knowledge