Episode 1: The 5 Whys, Simple, Powerful, and Easy to Get Wrong

Theo Ashworth

Episode 1: The 5 Whys, Simple, Powerful, and Easy to Get Wrong

Theo Ashworth

Picture a worker slipping on a puddle of oil on a plant floor. Cleaning up the spill deals with what is there now, but what allowed the oil to reach the floor, and why did it remain there? In this series we focus on investigating accidents, incidents and near misses, and on looking beyond the first explanation. Today we start with the 5 Whys, a method with roots on the Toyota factory floor. It can also be used for quality and production problems.

Nell Harding

And the method is almost insultingly simple. You ask why, then you ask why of the answer, and so on. I'll admit my first reaction is, surely that can't be a serious investigation tool.

Theo Ashworth

That's the strength and the trap, and we'll get to both. First, where it comes from. Sakichi Toyoda, who lived from 1867 to 1930, started out tinkering with looms. In 1924 his Type G automatic loom could stop itself the moment a thread broke. That idea became known as jidoka, building quality in by stopping at the first sign of a defect. And he had a habit, whenever something went wrong, of walking to the shop floor, looking at it himself, and asking why until the answer stopped being obvious. The phrase for that is genchi genbutsu, go and see for yourself.

Nell Harding

So at that stage it's more a temperament than a technique.

Theo Ashworth

Right. Taiichi Ohno, who lived from 1912 to 1990, took that temperament into the Toyota Production System and made it a discipline. One account says a line couldn't restart until the supervisor had walked to the spot and worked through the questions. Ohno is credited with the line that five whys equal one how. Ask why enough times, and the fix starts to show itself.

Nell Harding

Go on then, show us how it works. What about that oil spill?

Theo Ashworth

Here is an illustrative chain based on a published safety example. A worker slipped and fell. Why? There was oil on the floor. Why? Oil had leaked from a compressor. Why wasn't the leak detected? The compressor wasn't regularly inspected and repaired. Why not? It wasn't in the maintenance system. That gives us a possible underlying issue to investigate, not a conclusion to assume in every oil spill.

Nell Harding

And the point is what you'd have done at each stop along the way.

Theo Ashworth

Exactly. Stop at the oil and you might just clean the floor. Follow the evidence further and you may find a gap in inspection and maintenance that needs correcting. You would also need to ask why the spill remained on the floor. There may be more than one contributing cause, and the answers depend on the incident.

Nell Harding

Okay, but I can imagine a team doing this badly. What does a good session look like?

Theo Ashworth

It starts with a sharp problem statement. Instead of saying "safety issues in the plant," describe the event: a worker slipped and fell on oil on the plant floor. Add the time, location and other details when they are established. Then gather the people who saw it, plus someone to facilitate, and where it is safe, look at where the incident happened. The gemba again.

Nell Harding

And I'd guess the real discipline is in the answers.

Theo Ashworth

It is. Every answer needs evidence behind it, a log, a measurement, the physical part. If nobody knows, you stop asking and go and look. Guessing is how a tidy chain of whys leads you confidently to the wrong fix. And the five is a guideline. It might take three questions, it might take eight. You stop when you reach a cause you can act on, not when you hit a number.

Nell Harding

Here's where I'd push. Somewhere in that chain, someone says "the operator forgot," everyone nods, and the meeting ends. That's an answer. It's even true.

Theo Ashworth

And that's the most common place for the method to die. A quality writer at Scilife puts it well: labeling the root cause as human error doesn't put an end to the story, it is where the story begins. So you keep going. Why was that error possible? Was the procedure ambiguous? Was there any check that should have caught it? Why didn't a safeguard stop it?

Nell Harding

So you swap "why did the operator forget" for "why wasn't this checked."

Theo Ashworth

Yes, and there's a practical payoff beyond fairness. If people expect to be blamed, they stop reporting. Frame the questions around process, and the honest answers keep coming.

Nell Harding

Right, now the trap you promised. Because one chain of answers about an oil spill can look tidy. Most real incidents aren't.

Theo Ashworth

That's the big limitation. One chain is one corridor. Healthcare safety researcher Alan Card, writing in BMJ Quality and Safety, makes that point: complex failures branch into equipment purchasing, training, staffing, the physical design of a room, and a single chain walks past all of it. There's also the repeatability problem. Give two investigators the same incident and they can follow completely different paths, often steered by what they already suspected.

Nell Harding

Which is confirmation bias with a nice template.

Theo Ashworth

Even people inside Toyota have raised concerns. Teruyuki Minoura, a former managing director, has cautioned that it can turn shallow, or rest on off the cuff deduction, if it isn't grounded in what you've actually observed. So when a failure has several independent causes, you run a separate chain for each, or you reach for something built for branching, like a fishbone diagram or a fault tree.

Nell Harding

So the honest summary is that it's a brilliant first tool, and a dangerous only tool.

Theo Ashworth

I'd say so. For an accident, incident or near miss, go and look, ask why, check every answer against evidence, and when the chain forks, follow every branch. The same questions can help with quality and production problems too, but here our aim is to understand what happened and prevent people being harmed again.