Back to Blog
Root Cause Analysis

How to Write a D4 Root Cause That Survives Customer Review

What a D4 root cause must include, how to verify it, weak vs strong root cause statements, and the phrases that get an 8D sent back.

Vantage 8D TeamNovember 7, 20267 min read

Ask any customer quality engineer which part of an 8D they reject most often and you'll get the same answer: D4. The root cause section is where a report either earns trust or falls apart. A weak D4 leads to weak corrective actions, which leads to the same defect showing up again three months later with your name on it.

This guide covers how to write a D4 root cause that survives a customer review. It covers what to include, how to structure it, how to prove it, and the phrases that get a report sent straight back.

What a D4 root cause actually has to do

A good D4 answers three questions, and most rejected reports only answer the first one:

  1. Why did the defect occur? This is the occurrence root cause.
  2. Why didn't we detect it before it reached the customer? This is the escape root cause, sometimes called the non-detection root cause or, in Ford's Global 8D, the escape point.
  3. Why did our system allow both of those? This is the systemic root cause. Some customers ask for it explicitly in D4, others expect it to show up as D7 prevention actions.

If your D4 only explains how the part got made wrong, a seasoned reviewer will immediately ask how it got shipped. Answer that before they ask.

Step 1: Start from a solid D2

You can't find the cause of a problem you haven't defined. Before D4, make sure D2 tells you exactly what's wrong, on which parts, from which line, shift, lot and date, and how many. An Is / Is Not comparison is the single most useful input to root cause analysis, because every difference between the "is" and the "is not" is a clue.

If parts from press 6 crack and parts from press 5 don't, your cause almost certainly lives in whatever is different about press 6.

Step 2: Go wide before you go deep

Teams that jump straight to a 5-Why often pick the wrong branch and drill very confidently in the wrong direction. Go wide first.

An Ishikawa (fishbone) diagram forces the team to consider possible causes across six categories:

  • Man: training, shift handover, fatigue, staffing
  • Machine: wear, settings, maintenance, sensors, tooling
  • Method: work instructions, sequence, parameters, setup
  • Material: lot, supplier, storage, moisture, hardness
  • Measurement: gauge capability, inspection method, sampling plan
  • Environment: temperature, humidity, contamination, lighting

Once the fishbone is populated, use your D2 data to rule causes in or out. A cause that would affect all three shifts can't explain a defect that only happens on nights.

Step 3: Drill down with 5-Why, on both legs

Take the most likely branch and run a 5-Why. Then do it again for detection. A two-legged 5-Why looks like this:

Occurrence leg:

  1. Why did the housing crack? The boss was brittle.
  2. Why was it brittle? The PA66 was molded with excess moisture.
  3. Why was there excess moisture? The dryer ran at -10°C dew point instead of -40°C.
  4. Why? Desiccant regeneration failed and the dew point alarm was bypassed.
  5. Why was the alarm bypassed? Maintenance bypassed it during a sensor fault and there's no rule requiring alarms to be restored or time-limited.

Escape leg:

  1. Why did we ship cracked parts? Our outgoing check didn't find them.
  2. Why not? The check is visual and the crack only appears after a screw is inserted.
  3. Why don't we test insertion? It was never in the control plan for this part, because the PFMEA didn't list moisture as a cause of cracking.

Notice that each answer is something you can check, not an opinion. If you want help structuring this, our free 5-Why worksheet asks the next "why" for you based on your last answer. We also wrote a whole piece on why most 5-Why analyses fail.

Step 4: Verify the root cause

This is the step that separates a hypothesis from a root cause. Verification means you have evidence, not just a logical story.

The strongest form of verification is to turn the defect on and off:

  • Reintroduce the suspected cause and show the defect comes back
  • Remove the cause and show the defect disappears

Other acceptable evidence includes:

  • Lab analysis (material moisture, hardness, chemistry, fractography)
  • Data that correlates the cause with the defect in time, such as dryer logs matching the production window
  • Dimensional or capability studies before and after
  • A repeat of the escape: seed known-bad parts and show your current inspection passes them

Write the verification into the D4 itself, with numbers:

Parts molded from correctly dried material passed 50 of 50 insertion tests at 0.8 Nm. Parts molded from material at 0.35% moisture cracked in 6 of 50.

That one sentence does more for your credibility than a page of narrative.

Step 5: Write it up clearly

Here's a simple structure that works for almost any customer:

SectionWhat to write
Analysis methodTools used (Is / Is Not, Ishikawa, 5-Why, DOE, lab tests)
Occurrence root causeOne or two sentences, specific and technical
Occurrence verificationHow you proved it, with data
Escape root causeWhy existing controls missed it
Escape verificationHow you proved the control failure
Systemic causeThe process or system gap that allowed both

Keep each root cause statement short and specific. If you need a paragraph to explain the root cause, you probably haven't found it yet.

Weak vs strong root cause statements

WeakStrong
Operator errorTorque tool program P3 was selected instead of P5 because both programs share one unlabelled selector and the work instruction doesn't specify the program
Material issueCoil lot 4471 had hardness of 92 HRB against a maximum of 85 HRB; the supplier's certificate was not checked at receiving
Machine malfunctionPunch wear beyond 180,000 hits increased die clearance to 14% of material thickness; there is no hit-count-based PM trigger
Inspection missed itThe camera inspection window was 2 mm above the O-ring groove after a fixture change; recipes are not under change control

The strong versions name a specific mechanism, include a measurable fact and point naturally to a corrective action.

Phrases that get a D4 rejected

Avoid these, or at least don't stop at them:

  • "Operator error" or "human error." Humans make mistakes. The question is why the process allowed one. Ask what made the error possible and what should have caught it.
  • "Lack of training." Usually a symptom. Was the instruction unclear? Was the task error-prone? Did a change go uncommunicated?
  • "Root cause under investigation" submitted as a final D4.
  • "Supplier issue" with no data. If it's the supplier, show the evidence, and still explain why your incoming controls missed it.
  • Multiple unverified "possible" causes. If you list five possible causes and fix all five, the customer will assume you don't know which one it was. They'll be right.

Occurrence, escape and systemic causes in different 8D formats

Different customers use slightly different language for the same idea:

  • Standard 8D: root cause and non-detection cause, often both in D4
  • Ford Global 8D: root cause and escape point, with D5 actions required for both
  • VDA 8D: root cause of occurrence and root cause of non-detection, with emphasis on technical and systemic causes

We compare these in detail in Ford G8D vs VDA 8D vs standard 8D. Whatever the format, the reviewer is looking for the same thing: two verified causes and a clear link to the actions in D5.

A D4 checklist before you submit

  • Does D4 include both an occurrence and an escape root cause?
  • Is each root cause specific enough that the fix is obvious?
  • Is each root cause verified with data, ideally by turning the defect on and off?
  • Does the root cause explain everything in the Is / Is Not table?
  • Have you avoided "operator error," "training" and "supplier issue" as final answers?
  • Does every root cause have a matching action in D5?

If you can tick every box, your D4 will hold up in front of almost any customer. If you write a lot of these, Vantage 8D builds an Ishikawa diagram from your D2, D4 notes and past 5-Why analyses, then polishes the final wording so your evidence reads as clearly as it deserves to.

Try AI-Powered 8D Reporting Free

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