Monday, 7:40 a.m. The customer writes: 340 bearing bushings from the last delivery are oversized on the outer diameter, and the assembly line has stopped. Attached is a form with eight empty fields and one line: "Please send the 8D report by Friday."
If you work in quality at a supplier, you know this email. In this article we walk through an 8D report from start to finish using exactly this case. At the end you'll find a matching template for free, as Excel and as a printable PDF. The Excel file also includes this example as a filled-in sheet.
What is an 8D report?
An 8D report is how a supplier shows a customer that a quality problem has been solved. "8D" stands for eight disciplines: eight working steps from forming the team to closing the case. The method comes from Ford (late 1980s) and is standard in the automotive industry today, where the German VDA has published its own volume on it. Mechanical engineering, medical devices and electronics use it too.
The method and the report are not the same thing. The 8D method is how you work the problem. The 8D report is the document that records that path. A report written on Friday from memory often looks complete and still helps nobody. It is meant to be a worksheet that grows while you work.
Some variants, such as Ford's "Global 8D", add a step D0 in front: emergency response and the question of whether the problem needs a full 8D at all. For most customer complaints, D1 to D8 are enough.
The eight disciplines at a glance
| Step | Content | Key question |
|---|---|---|
| D1 | Form the team | Who knows the product and process, and who can decide? |
| D2 | Describe the problem | What exactly is wrong, where, since when, how many? |
| D3 | Containment | How do we protect the customer until the cause is known? |
| D4 | Find the root cause | Why did the defect occur, and why wasn't it detected? |
| D5 | Choose corrective actions | What removes the cause, and how do we prove it? |
| D6 | Implement and verify | Is the action in place, and does the data show it works? |
| D7 | Prevent recurrence | Where else could the same defect happen? |
| D8 | Closure | Does the customer agree, and what does the team take away? |
Many customers expect D1 to D3 within 24 hours, and the full report with root cause and corrective action a few days to a few weeks later. The exact deadlines are in your supplier quality agreement.
A case from the turning shop
The case we'll work through: a supplier runs five CNC lathes making bearing bushings for an electric motor manufacturer. Nominal outer diameter: Ø 22 mm, tolerance h7, so 22.000 to 21.979 mm. At assembly, bushings won't press into the housing. The customer measures and finds parts up to 22.015 mm.
The example is made up, but typical.
D1: Form the team
8D is teamwork. Whoever fills in the form alone at a desk usually finds the cause they suspected before they started.
The team needs people who really know the product and the process, plus someone who can approve actions:
- Team lead: quality manager
- Production: turning shop shift supervisor, an operator from the affected shift
- Metrology: measuring room
- Customer contact: key account manager from sales, so commitments to the customer are aligned
In practice these are usually the people who already worked on the incident together. They know what happened when and what has already been tried. If nobody among them can approve actions or handle metrology, bring that person in.
D2: Describe the problem
D2 decides whether D4 will work. A fuzzy description leads to a fuzzy cause.
Bad: "Parts out of tolerance."
A good description is measurable and also says what is not affected. The is/is-not analysis has proven itself here:
| Is | Is not | |
|---|---|---|
| What | Outer diameter 22.008 to 22.015 mm (tolerance 21.979 to 22.000) | Inner diameter, length, surface |
| Where | Machine DZ-3 | Machines DZ-1, DZ-2, DZ-4, DZ-5 |
| When | Late shift on 14 Oct, from about 7:30 p.m. | Early shift on 14 Oct, all other days |
| How many | 340 of 2,000 parts in lot 24-117 | Other lots |
This table already narrows the search a lot: one machine, one shift, one point in time. What happened at DZ-3 at 7:30 p.m. becomes the central question.
D3: Containment
Containment protects the customer while the cause is still unknown. That doesn't solve the problem, it only contains it: defective parts stop reaching the customer.
In the example:
- All parts from lot 24-117 blocked: in stock, in shipping and at the customer
- 100% inspection of the outer diameter on all blocked parts, sorting
- Replacement delivery of inspected parts within 24 hours
- Until closure: every delivery from DZ-3 gets 100% inspection and marking
Then comes the part that's often missing: is the containment working? Here it is, the replacement delivery went through the customer's assembly without findings. That goes into the report with a date.
D4: Find the root cause
This is where the report earns its keep. A good D4 answers two questions separately:
- Why did the defect occur? (occurrence cause)
- Why wasn't it detected before it reached the customer? (escape or non-detection cause)
In the example, the 5 Why method takes the first question this far:
- Why is the diameter too large? The tool offset was wrong after an insert change.
- Why was it wrong? The offset was entered on tool T03 instead of T04.
- Why was that possible? T03 and T04 use the same insert, and the control's offset table shows numbers only, no description.
And the second:
- Why wasn't the defect detected? There was no first-piece inspection after the insert change.
- Why not? The inspection plan only requires it after a tool change, and an insert change doesn't count as one.
- Why? The inspection plan dates from a time when complete tool holders were swapped. Since inserts are now changed directly in the machine, it no longer fits.
A good D4 needs one more thing, and that's proof. The next day the team recreated the situation at DZ-3, deliberately put the offset on T03 and turned one part. Result: 22.012 mm. The cause is now confirmed, not just suspected.
How to recognise a good D4:
- It is technically specific (tool, value, inspection plan) instead of "production error".
- It explains the pattern from D2: why only DZ-3, why from 7:30 p.m.
- It covers both the occurrence cause and the escape cause.
- It is confirmed by a test or data.
- It ends at something you can change: inspection plan, display, procedure. It never ends at a person.
For the search itself, besides 5 Why, an Ishikawa diagram (man, machine, material, method, measurement, environment) helps, and so does the plain comparison: what was different at DZ-3 at 7:30 p.m.? The template has its own sheet for the 5 Why chains.
D5: Choose corrective actions
Corrective actions target the cause from D4. Every cause needs at least one action, and before it is rolled out you should have evidence that it works.
| Cause | Corrective action | Evidence before rollout |
|---|---|---|
| Offset could be entered on the wrong tool | Tool description stored in the control's offset table (T04 "OD finish Ø22") | Trial run: operators find the right tool without asking |
| No first piece after insert change | Inspection plan: first-piece inspection after every insert change, sign-off in the machine log | Trial run over two shifts |
What's missing from this list: "operator retrained". Training doesn't change the process that allowed the defect. In the automotive industry it is usually not accepted as the only corrective action. It can support, but it can't carry the fix.
D6: Implement and verify
D6 has two parts, and the second one tends to be forgotten. Implementation gets a date and an owner:
- Tool descriptions stored at DZ-3: 21 Oct, turning shop shift supervisor
- New inspection plan effective 21 Oct, quality manager
Whether the action works only shows in the data after rollout. In the example: ten lots from DZ-3 without a single part out of tolerance, with outer diameter readings stable in the middle of the tolerance band. Only then is the 100% inspection from D3 lifted, and that belongs in the report too.
D7: Prevent recurrence
D6 fixes this one case. D7 asks where else the same defect could happen. In the example:
- Tool descriptions stored on all five lathes, not just DZ-3
- The inspection plan template for new parts updated: an insert change counts as a tool change
- Process FMEA extended: failure mode "offset entered on the wrong tool" assessed
- Lessons learned passed on to the second plant
Leave D7 empty and the same defect shows up again in six months as a new complaint, just on a different machine.
D8: Closure and recognition
The customer confirms closure and the report is filed. And the team gets recognition. That sounds like a formality, but it is why people will say openly what really happened at the machine during the next 8D. If all you look for after a complaint is someone to blame, the next round of answers will be polished.
Typical mistakes in 8D reports
Most reports that customers send back fail in the same places:
- D2 too vague: "Parts defective" instead of dimension, quantity, location and time. Without is/is-not the search stays wide.
- Containment confused with corrective action: Sorting and 100% inspection belong in D3. If they show up in D5, the cause hasn't been removed.
- "Human error" as root cause: What matters is why the process let the defect through. Knowing who made it barely helps with the corrective action.
- Only the occurrence cause: Why the defect wasn't detected is missing, so the next defect slips through just as quietly.
- Cause assumed, not proven: "Probably tool wear" is a hypothesis. A hypothesis doesn't belong in D4 yet.
- D6 with a date only: "Implemented on 21 Oct" doesn't show that it works. The customer wants to see measurements from after the rollout.
- D7 forgotten: The fix applies to one machine, and the four identical ones next to it stay as they were.
- Report written after the fact: Times, measurements and decisions get reconstructed from memory, and the very details that D2 and D4 depend on are lost.
Free 8D report template
The template follows the eight disciplines as described in this article:
- Sheet "8D Report": header (complaint number, customer, part, lot) and D1 to D8 with fields for owners, dates and proof of effectiveness
- Sheet "5 Why": separate chains for the occurrence cause and the escape cause
- Sheets "Example" and "Example 5 Why": the bearing bushing case, fully filled in
Download: 8D report template as Excel (.xlsx) or as a printable PDF. No sign-up, no email address required.
If your customer prescribes their own form, use theirs. The content is the same, and with the template you have it all together before you copy it over.
When the report writes itself from the timeline
The last mistake on the list is the most common one. An 8D usually gets written at the very end, and by then the note someone made at the machine at 7:30 p.m. on 14 October is missing, or it's scattered across emails, tickets and Teams chats.
That's exactly why we built IncidentLive. During an incident, tickets, measurements, notes and decisions land on one shared timeline. From it, IncidentLive drafts D1 to D8, and every statement links back to the entry it came from. Your team still finds the root cause, and a human always gives the approval. The page 8D reports with IncidentLive shows how it works in detail.