Back to Blog
Root Cause Analysis

Why Most 5-Why Analyses Fail (And How to Fix Yours)

The five most common 5-Why failure patterns, from stopping at human error to skipping the system level, and a simple framework to go deeper.

Vantage 8D TeamNovember 7, 20266 min read

The 5-Why is the most used root cause tool in manufacturing, and probably the most misused. It looks simple: ask "why" five times and you'll reach the root cause. In practice, most 5-Why analyses stop too early, follow the wrong path, or end in a conclusion nobody actually checked.

This article covers why most 5-Why analyses fail, the five failure patterns we see again and again, and a simple way to make yours hold up in front of a customer or an auditor.

Why the 5-Why is worth getting right

The 5-Why comes from the Toyota Production System. Its strength is that it forces you past the obvious answer. Its weakness is that it has almost no built-in guardrails. A fishbone diagram makes you consider six categories. A DOE makes you control variables. A 5-Why only asks you to keep asking, so the quality of the answer depends entirely on the discipline of the person asking.

When a 5-Why fails, the corrective action fails with it. That's how you end up writing an 8D for the same defect twice.

Failure pattern 1: Stopping at "human error"

This is the most common failure by far. It goes like this:

  1. Why was the wrong label applied? The operator picked the wrong roll.
  2. Why? The operator made a mistake.

Root cause: operator error. Corrective action: retrain the operator.

The problem is that humans make mistakes all the time. A process that relies on nobody ever making one is a fragile process. When you hit a human error, that's not the end of the analysis. It's a signal to ask two better questions:

  • What made the mistake possible? Two label rolls that look identical, stored side by side, with no scan check.
  • What made the mistake likely? A changeover in the last hour of a shift, with the label part number printed in 6-point text.

Keep going until you reach something you can change in the process, not in a person.

Failure pattern 2: A vague problem statement

A 5-Why can only be as good as the problem it starts from. "Parts are failing" isn't a starting point. Neither is "quality issue on line 2."

Compare:

  • Vague: Why are housings cracking?
  • Specific: Why did 14 of 2,000 PA66 housings from press 6, second shift, 3 November, crack at the screw boss during insertion at 0.6 Nm?

The specific version already rules out half the possible causes. It also tells you where to go and look. Spend five minutes sharpening the problem statement before you ask the first "why." If you're working inside an 8D, that's your D2.

Failure pattern 3: Answers that are opinions, not facts

Every answer in a 5-Why should be something you checked, not something you believe. A typical failing chain:

  1. Why did the bolt loosen? It wasn't torqued enough.
  2. Why? The torque tool was probably out of calibration.
  3. Why? Calibration is probably overdue.

Two "probablys" in a row means the analysis has left the shop floor and moved into the meeting room. Before you write the answer to each "why," ask: how do we know? Then go and look. Pull the calibration record. Check the torque tool's last readings. Measure the residual torque on suspect bolts.

A quick sanity test: read your chain backwards using "therefore." If the calibration was overdue, therefore the tool read low, therefore the bolt was under-torqued, therefore it loosened. Each "therefore" should be supported by evidence, not assumption.

Failure pattern 4: Only one leg

Most 5-Why analyses explain why the defect was made. Very few explain why it wasn't caught. Customers notice.

A complete analysis has at least two legs:

  • Occurrence leg: why did the defect happen?
  • Detection leg: why did our controls let it escape?

Many teams add a third:

  • System leg: why did our management system allow both? Think missing FMEA entries, change control gaps, or maintenance plans that don't cover a critical component.

If you only fix the occurrence leg, you've fixed one way the defect can happen. Fixing the detection leg protects you against the ways you haven't discovered yet.

Failure pattern 5: Going too far, or too sideways

There's an opposite failure too. Some 5-Why chains keep going until they reach something nobody can act on:

  1. Why was the PM missed? Maintenance was short-staffed.
  2. Why? The budget was cut.
  3. Why? Sales were down.
  4. Why? The economy.

The useful stopping point is the last cause you can actually control and change, usually a process, standard or system inside your plant. "Five" is a guideline, not a rule. Some problems need three whys. Some need seven.

The sideways version is when the chain changes subject halfway through. If "why" number three suddenly talks about a different part or a different shift, you've jumped branches. Go back and follow one thread at a time.

A simple framework that works

You don't need a new tool. You need a few rules applied every time:

  1. Start from a fact-based problem statement with what, where, when and how many.
  2. Go wide first. Sketch an Ishikawa diagram (man, machine, method, material, measurement, environment) and use your data to pick the most likely branch.
  3. Run two legs: occurrence and detection. Add a system leg for serious or repeat problems.
  4. Verify every answer. If you can't verify it, mark it as a hypothesis and test it.
  5. Never stop at a person. Ask what made the error possible and likely.
  6. Stop at the last controllable cause.
  7. Check backwards with "therefore."
  8. Prove the root cause by turning the defect on and off where you can.

A worked example

Problem: 3 of 400 wire harnesses returned by the customer with a terminal pushed back out of the connector housing (line 4, night shift, week 41).

Occurrence leg:

  1. Why was the terminal pushed back? The terminal's locking lance wasn't engaged.
  2. Why? The terminal wasn't inserted fully. Measured insertion depth on returned parts was 1.2 mm short.
  3. Why? The insertion was done by hand, and the operator relies on feeling a "click."
  4. Why wasn't the click felt? On night shift the station uses a new batch of housings with a slightly stiffer lance, so the click is much softer (confirmed with a force gauge: 8 N vs 14 N on the old batch).
  5. Why wasn't the method adjusted? Housing lot changes aren't treated as a process change, so nobody reviewed the effect on the manual operation.

Detection leg:

  1. Why did we ship it? The pull test didn't catch it.
  2. Why? The pull test is a sample of 1 harness per hour, at a force that a partially engaged lance can still hold.
  3. Why wasn't there a 100% check? The PFMEA rated "terminal not locked" as low occurrence because the click was considered reliable.

Verification: with the new housing batch, 9 of 40 trial insertions by three operators were short. Adding a secondary lock (TPA) closure check caught all 9.

Notice there's no "operator error" in this chain. The operator did exactly what the process asked. The process changed under them.

Making 5-Why easier on your team

The fastest way to improve your 5-Why analyses is to make the right questions automatic. Our free AI 5-Why worksheet asks each next "why" based on your last answer and pushes back when you stop at human error. Inside an 8D, Vantage 8D also uses your 5-Why to build the Ishikawa diagram in D4.

For the next step after the 5-Why, read our guide on how to write a D4 root cause, and for the wider list of traps, see common mistakes in root cause analysis.

Try AI-Powered 8D Reporting Free

Generate complete 8D reports, run AI-guided 5-Why analyses, and evaluate supplier 8Ds — all in one place.