Five Whys: root cause analysis

A technique for finding the cause of a problem by asking "why?" five times in a row, each time about the answer you just gave. In the classroom you work in pairs: the problem is on the screen, the pair asks "why does this happen?" and writes the answer down, and then asks why about that answer. After five questions there is a chain on the sheet, from the problem at the top to the cause at the bottom. Then the class compares which cause each pair found.

With the default settings the whole structure takes ± 11 minutes, briefing and look-back included.

With the default settings about eleven minutes for one problem: a minute to form the pairs, five whys of a minute each and two minutes comparing. That is eight minutes of steps; the preparation adds a minute for reading out the problem and two minutes of reflection at the end. Every further problem takes seven minutes, because the pairs stay seated: two problems take about eighteen minutes. With three whys you are done in nine minutes, with seven in thirteen. You set the times yourself, in steps of fifteen seconds (all the whys together in one row), and the player moves on to the next why by itself.

No. Five is the name of the structure, not a law of nature. In the preparation you choose between three and seven: three in a short lesson or with young students, seven for a stubborn problem. What is fixed is that every question is about the previous answer. Expect it to get hard around the third why. That is not a sign to stop, but the point where the digging begins.

A good problem is a situation you can see happening, that the class recognises, and that has more than one possible cause. "Homework is not handed in on time" works; "students are lazy" does not, because that is already a judgement, and neither does "the printer is broken", because that cause is known after one question. Leave names out and phrase it as one short sentence. No problem in mind? Choose "AI writes them" in the preparation: you get a list of recognisable problems for your subject and the age of your class, and keep what fits. If you prefer your own assistant, such as Microsoft Copilot, copy the request from the same screen.

In pairs because then all students dig at the same time. In a class discussion one student digs and the rest listen; here twelve pairs ask their questions at once, and each keeps the other sharp when they fall back on the problem. On one sheet because the chain is the work: five answers that build on each other are more than you can keep track of in your head, but written one under the other you read them back in one go. With two sheets you get two chains, and at the end there is nothing to compare together.

That is exactly the yield. In the last step each pair says in one sentence what is at the bottom of its chain: only the bottom line, not the whole chain. If you hear different causes for the same problem, ask about the difference: where did you go a different way? That is how the class learns that one problem can have more than one cause. Close with the question of which cause is within your own reach; that is where what you do differently tomorrow begins.

It is the same ladder, the other way round. Five Whys digs downwards: why does this problem happen? At the bottom is a cause. Nine Whys climbs upwards: why is this important to you? At the top is a purpose. In Nine Whys one asks and the other talks, then they swap roles, and nothing is written down. It is a conversation about what drives you, not an analysis of a problem.

Both are about what goes wrong, but they look at it differently. Five Whys takes one problem and looks for the cause. The worst plan takes a goal, works out how you would be sure to ruin it, and then asks which of those things you are secretly doing already, and which one thing you will stop. If you know what goes wrong and want to understand why, take Five Whys. If the group feels it is getting in its own way but cannot see how, The worst plan works better.

Beforehand: phrase the problem as a situation, without names, and have one sheet ready per pair. At the start, say that every question is about the previous answer and that it is allowed to get hard. During the questions: walk round and read along. If every line answers the line above it, it is going well; if they are five answers to the problem itself, point at the previous answer and ask: why does that happen? Give no causes yourself. When the class compares: ask about the difference between chains, not for the right answer, and end with what is within your own reach.

From Japanese industry. Sakichi Toyoda, the founder of what later became Toyota, already used the question; Taiichi Ohno made it a fixed part of the Toyota Production System and described it in his book on that system (1978). His well-known example: a machine stops because a fuse blows, and after five whys the cause turns out to be a missing strainer on the lubrication pump, which let metal scrap get into the pump. Through lean the technique also reached healthcare, government and education. The LudensFlow version turns it into a classroom structure: pairs digging at the same time, one sheet per pair, and a closing step in which the class compares the chains, because in the classroom the gain is in the difference between the chains, not in one right cause.