Academy › Practical guides › Quality tools
How to write an 8D report a customer will actually accept
Most 8Ds get rejected at D4 — not because of formatting, but because the root cause is a person's name.
Updated 13 November 2026 · 8 min read · Written for working labs and quality teams
Short answer
An acceptable 8D proves four things: the problem is quantified (D2), all affected material is contained with traceable quantities (D3), the occurrence cause and the escape cause are separately established and evidenced (D4), and the permanent actions are verified and propagated into the PFMEA, control plan and work instructions (D5–D7). Most rejections happen at D4, where "operator error / retraining" is offered as a root cause instead of a system weakness that allowed the error.
On this page
D0 to D8, with what the customer checks
| Discipline | What goes in | What the customer checks |
|---|---|---|
| D0 Emergency response | Immediate action and who authorised it | Was anything done before the investigation started? |
| D1 Team | Cross-functional team, owner, sponsor | Does the team include process, quality and engineering — not just quality? |
| D2 Problem description | 5W2H, defect rate, part numbers, dates, trend, photos | Is the problem quantified and bounded, or described as 'some parts were bad'? |
| D3 Interim containment (ICA) | Sorting/segregation results for stock, WIP, in transit and at customer | Are quantities reconciled? Was the containment itself verified? |
| D4 Root cause | Occurrence cause AND escape cause, with evidence (5-Why + data) | Would the cause explain why the existing controls failed? |
| D5 Permanent corrective action (PCA) | Chosen action, alternatives considered, validation evidence | Does the action remove the cause, and was it verified before rollout? |
| D6 Implementation & validation | Rollout, removal of ICA, before/after data | Is the defect rate actually lower, with a defined monitoring window? |
| D7 Prevention of recurrence | Systemic read-across: PFMEA, control plan, work instructions, similar lines/parts | Was the same failure mode made impossible elsewhere? |
| D8 Closure | Documented closure, evidence pack, team recognition | Is the evidence pack complete enough to survive a later audit? |
The two causes you must separate
Occurrence cause: why the defect was made. Escape cause: why the existing controls did not catch it before shipment. A missing clip because the fixture wore out is an occurrence cause; the fact that the end-of-line check does not verify clip presence is an escape cause. An 8D with only one cause is normally returned, because fixing one leaves the other route open.
For each cause, run a 5-Why chain and stop only at a level where a controllable system element appears — maintenance interval, error-proofing, inspection method definition, supplier change control. Then verify the chain with evidence: the worn fixture's measurement, the work instruction that lacks the check, the audit record that never covered it. A 5-Why chain with no attachments is a story, not an analysis.
Practical note. "Operator error" is not a root cause; it is the last visible link in a chain. Ask why the method allowed the error: no poka-yoke, ambiguous instruction, rushed cycle time, inadequate training verification, or a temporary worker on an unqualified station. Fixing the system element closes the chain; retraining alone leaves it exactly as it was.
Containment is a quantity, not a sentence
- List every population: finished goods, WIP, in transit, customer stock, returned parts, and parts already assembled into higher-level products.
- For each, state the quantity, the sort method, the number of defects found, the date/time, and who did it. Reconcile the numbers so they add up to the shipped quantity in D2.
- Identify the containment with a traceable mark or label so an auditor can tell sorted from unsorted stock three months later.
- Verify the containment itself (e.g. re-check a sample of the sorted lot) and only remove it after D6 validation data shows the PCA works.
D7 is where the 8D pays for itself
Once the PCA is validated, update the PFMEA (new prevention/detection control, re-scored S/O/D and Action Priority), the control plan (control method, sample size, frequency, reaction plan), the work instruction and the training/qualification records. Then perform read-across: which other part numbers, lines or suppliers share the same failure mode? A customer asking "why did this happen again on another line six months later" is asking about D7, not D4.
Keep the evidence pack together — problem description, containment records, cause-analysis data, validation results, updated documents with revision numbers, and closure approval. An 8D that cannot be reconstructed from its attachments will not survive a subsequent customer or certification audit.
Where this goes wrong in practice
Naming "operator error", "lack of awareness" or "human mistake" as the root cause and closing with retraining.
Writing D3 as "all stock checked" with no quantities, no reconciliation and no way to prove which stock was checked.
Skipping the escape cause entirely, so the same defect can ship again undetected.
Choosing a PCA before the root cause is confirmed, and never validating whether it reduced the defect rate.
Leaving the PFMEA, control plan and work instructions untouched at D7 — then treating the next occurrence as a brand-new problem.
Frequently asked
Is 8D required by ISO 9001 or IATF 16949?
No — 8D is a customer-driven problem-solving method, widely required through automotive customer-specific requirements. The standards require corrective actions, root-cause analysis, evaluation of effectiveness and updates to the QMS; 8D is one structured way to document that work.
What is the difference between D3 containment and D5 permanent action?
D3 is a temporary measure that protects the customer immediately (sorting, extra inspection, quarantine); it does not change the process. D5 changes the process so the cause cannot recur. D6 removes the temporary measure only after the permanent one has been validated with data.
How long should an 8D take?
Customers usually set their own cadence — often an initial response within 24 hours, containment evidence within a few days, and a final 8D within 10–30 days. Ask for the customer's specific timing requirement in their supplier manual rather than assuming a fixed number.
Do I need both a 5-Why and a fishbone diagram?
No method is mandated. What matters is that the analysis separates occurrence from escape, is supported by evidence, and reaches a controllable system cause. Many teams use a fishbone to structure candidate causes and 5-Why to drill down on the confirmed ones.
Structure the whole thing
Build the 8D with editable D0–D8 sections and an exportable report
Work through each discipline with prompts for containment quantities, occurrence vs escape causes, PCA validation and D7 document updates — then export the assembled report.
Open the free 8D problem-solving tool →Runs in your browser. Export the report free. No sign-up needed.
LabKitPro guides are written from published standards and accreditation-body documents — references, clause numbers and paraphrased intent only, never verbatim standard text. They are for preparation and study, not a certification or accreditation decision, and not legal advice. ISO standards are trademarks of ISO; IATF 16949 is published by the International Automotive Task Force and is not an ISO standard. The current laboratory standard is ISO/IEC 17025:2017; there is no 2025 edition. Global ACI replaced IAF and ILAC on 1 January 2026.
Academy by standard · All practical guides · Free browser tools