From Toyota's production system, credited to Sakichi Toyoda and first set down by Taiichi Ohno (Toyota Production System, 1978).
Asks why a problem happened, then asks why again of each answer, producing a short chain that ends at a cause in the process worth fixing.
Five whys on a late order, with the choice at every turn: stop and fix, or ask again
At each of the ops manager's turns, choose: fix the last answer, or ask why again. The verdict says whether that answer was still a symptom. The cause worth fixing is the first one that names how the process is built. Illustrative.
Ops manager
Warehouse team
Ops manager
Warehouse team
Right calls: 0 of 6Right calls: 1 of 6Right calls: 2 of 6Right calls: 3 of 6Right calls: 4 of 6Right calls: 5 of 6Right calls: 6 of 6
- The problem (Warehouse team). A customer's order of forty pallet racks shipped nine days after the promised date, and the customer wants to know what will stop it happening again.
- Why 1 (Ops manager). The only thing known so far is the problem itself.
- Stop: apologize and send the order by courier The customer gets the racks and a credit. A patch: this order is handled, and nothing stops the next one.
- Ask: why did the order ship late? The team checks the order history before answering. Right: so far there is only an event, no cause.
- Answer 1 (Warehouse team). The order sat waiting for the uprights, the tall end frames, which were not on the shelf when it came in.
- Why 2 (Ops manager). A stockout is a cause of the delay, and it has a cause of its own.
- Ask: why were the uprights out of stock? The team pulls the stock record for the item. Right: an empty shelf is still a symptom.
- Stop: rush-order more uprights Purchasing pays for a rush delivery. A patch: the shelf is full again until the next time it empties.
- Answer 2 (Warehouse team). The stock record shows the uprights hit zero two weeks before the order, and no purchase order followed.
- Why 3 (Ops manager). Nobody reordered. It is tempting to stop at the person.
- Stop: tell the buyer to watch stock more closely The buyer is asked to be more careful. Blame: the buyer works from a report, and the report has not changed.
- Ask: why did no one reorder? The buyer explains how reorders are triggered. Right: a person's miss usually has a process behind it.
- Answer 3 (Warehouse team). The buyer reorders from a weekly low-stock report. The uprights never appeared on it.
- Why 4 (Ops manager). The report missed an item. Checking by hand would catch it, at a cost.
- Ask: why did the report miss the uprights? The team looks at how the report chooses items. Right: the report is part of the process, so look inside it.
- Stop: have the buyer check every item by hand each week Four hours a week go into the check. A workaround: it costs time and still misses things.
- Answer 4 (Warehouse team). The report lists only items whose stock is below their reorder point. The uprights had no reorder point at all.
- Why 5 (Ops manager). One item is missing a setting. Is it the only one?
- Stop: set a reorder point for the uprights The uprights now appear on the report when they run low. Close: it fixes one item and leaves the gap that caused it.
- Ask: why did the uprights have no reorder point? The team checks how items are set up in the stock system. Right: a missing setting on one item may be missing on many.
- Answer 5 (Warehouse team). Items added to the stock system have no reorder point unless someone types one in, and nothing requires it. 212 items have none.
- Stop or ask again (Ops manager). The answer is now about the design of the process, and it can be changed.
- Stop: make the reorder point a required field and fill in the 212 Every new item now needs a reorder point before it can be saved. The cause to fix: it names the process and closes the gap for every item.
- Ask: why does the setup screen allow it? Because nobody thought about reordering when the screen was built. Too far: the answers now are about attention, and give nothing to change.
An example
A shelving wholesaler shipped a customer's order of forty pallet racks nine days late, and the customer asked what would stop it happening again. The operations manager took twenty minutes with the warehouse lead and the buyer.
Why was it late? The uprights were out of stock. Why? They had run out two weeks earlier and nobody reordered. It would have been easy to stop there and tell the buyer to be more careful. Why did nobody reorder? The buyer works from a weekly low-stock report, and the uprights never appeared on it. Why not? The report lists only items below their reorder point, and the uprights had none. Why? Items added to the stock system get no reorder point unless someone types one in, and nothing requires it.
That answer named how the process was built, so they stopped. A query found 212 items with no reorder point. The fix was a required field on the setup screen and an afternoon filling in the 212. Stockouts on recently added items fell to almost none over the next quarter.
Why it works
Each why moves the question from what happened to what allowed it to happen. The first answers describe the event: the shelf was empty. Later answers describe the system: new items are set up without a reorder trigger. A fix at the level of the event repairs one order; a fix at the level of the system repairs every order like it.
The method's real content is its stopping rule. Stop at the first answer that names how the process is built and that someone can change. Stopping earlier treats a symptom. Going further reaches answers about attention or human nature, which give nothing to change. The number five is a rule of thumb for going one or two steps past the point where the answer first feels sufficient, which is usually where the person, not the process, is being blamed.
The thinking behind it
Sakichi Toyoda Sakichi Toyoda, the founder of the Toyota group, built power looms that stopped themselves when a thread broke or ran out, the origin of jidoka: building quality in by halting at the first defect. Toyota and Taiichi Ohno credit him with the habit of asking why five times. No text of his describes the method, so the credit rests on the company's own account of where its practices began.
Taiichi Ohno Taiichi Ohno, the Toyota executive regarded as the father of the Toyota Production System, gave the first printed account in his Toyota Production System (1978; in English, 1988). He called asking why five times the basis of Toyota's scientific approach, and traced the practice to Sakichi Toyoda. The method has its critics, including a former Toyota executive, who call it too shallow to find root causes reliably, since different investigators often reach different answers.
How to run it
- State the problem as a fact. What, when, and how much: the order shipped nine days late.
- Bring the people who do the work. The warehouse lead and the buyer know answers a manager would only guess at.
- Ask why and write down one answer. Each answer should be a single direct cause that someone can check.
- Check each answer before going on. Look at the stock record or the report; a wrong answer sends the chain down the wrong path.
- Ask why of the answer. Keep going while the answer still describes an event or a person's slip.
- Stop at the process. Stop at the first answer that names how the process is built and can be changed.
- Fix it and watch for a return. Make the change, then check whether the problem comes back over the next weeks.
Where it goes wrong
The chain stops at a person. The buyer forgot, so the buyer gets a warning, and the report that failed the buyer stays as it was. Whenever an answer names a person, ask what let the mistake happen.
One chain is drawn where there are several causes. Many problems have two or three causes at once, and a single chain finds one of them. Spot it when the fix is in place and the problem returns; draw the causes by category instead.
Answers go unchecked. Each guess becomes the premise of the next question, and the chain ends at a confident wrong answer. Check every answer against a record before asking why again.
When to use something else
- Fishbone Diagram. When the problem probably has several causes in different parts of the work, and one chain would miss most of them.
- Blameless Postmortem. When an incident needs a full written review that a whole team can learn from without anyone being blamed.
Where it sits in SOLVE eX
Go to the source
- Toyota Production System: Beyond Large-Scale Production, Japanese edition 1978; English edition, Productivity Press, 1988, 1978
- The Toyota Way, 2004