8D Report Example: A Worked Customer Complaint

An 8D report is the structured record a supplier sends a customer after a quality complaint. It names a team, describes the problem, contains suspect product, finds the root cause and the escape cause, proves the fix works and stops the problem returning. This guide fills one in end to end with invented numbers.

One fact about our own offer, so you can judge this fairly. The EDWartens quality engineer job simulation is a training project at a simulated company, done day by day with project documents, emails and tickets. It is not employment, an internship or a work placement. The first week is free and every ticket is checked as you hand it in. A graded, verifiable certificate is optional, and the longer Premium version is paid; prices show in your own currency before you pay. It shows on your EDWartens profile under Work experience, marked as a job simulation.

When does a customer ask for an 8D?

A customer asks for an 8D when a part has failed in their plant or in service and they want proof that you understand why and that it will not happen again. In automotive supply the request usually comes through a supplier quality engineer, and the customer-specific requirements in your agreement set the clock. The wider framework is the quality management system: ISO 9001 clause 8.7 covers control of nonconforming outputs, and clause 10.2 covers nonconformity and corrective action. Automotive customers add their own rules on top through IATF 16949. The ASQ page on 8D gives the standard outline of the method.

One thing separates a good 8D from a poor one more than any other: the good one is written from evidence, step by step, while the poor one is written backwards from the answer somebody already wanted. Everything below is built so each step has something you can check.

The eight disciplines at a glance

Treat D0 as optional paperwork only if your customer says so. In practice it is the step that stops you wasting the first day.

The invented case

Everything here is an invented teaching example. The supplier is Thornby Castings and the customer is Ashgrove Drivetrain, both made up, and the numbers are chosen to make the method clear, not copied from any real company.

Thornby makes aluminium pump housings, part PH-4417, by high-pressure die casting. The drawing says the sealing face must be flat within 0.060 mm. One Tuesday Ashgrove's end-of-line leak test rejects 17 housings from a delivery of 1,920. Their quality engineer measures three and finds flatness of 0.071, 0.083 and 0.094 mm. They raise a complaint with the lot number and ask for containment in one working day and an 8D in ten.

D0 to D3: team, problem and containment

D1, the team. A quality engineer leads. The team is a casting process engineer, a die maintenance technician, a shift supervisor, a gauge room technician and a purchasing buyer, because purchasing turns out to matter. Each person has a role and a date they joined.

D2, the problem description. Use is and is-not. The words below are the kind of statement a reviewer wants to see:

The second row is the first clue: one cavity of one die, starting on a date. A problem that has a where and a when already has a short list of causes.

D3, containment. Containment means no suspect part reaches another customer, and it comes before anyone knows the cause. Make a stock table at once, counting every location where the lot or later lots could be.

The totals add up to the delivery, 1,920. If a containment table does not add up to the number shipped, you do not know where the suspect parts are. Add a rule to the table: certified clean parts get a coloured label for the next four weeks so a good part is never confused with a sorted one.

D4: root cause and escape cause

Start with data, not opinion. A Pareto chart of the defect log for the past three weeks shows where the scrap really is:

Flatness is the largest bar, and a short capability study on cavity B parts cast that week gives a Cpk for flatness of 0.92 from 110 parts, against a customer requirement of at least 1.50 in series. A Cpk below 1.00 means the process cannot hold the tolerance even when it is stable. The NIST handbook explains capability and control charts in more detail.

The team now walks the 5-why chain, and every answer must be proved with something you can see or measure:

1. Why did the housings leak? Flatness was 0.071 to 0.094 mm against 0.060 mm maximum. Proof: gauge readings on retained parts.
2. Why was the face not flat? The casting sticks to the lower die half on ejection and distorts as the ejector pins push. Proof: drag marks on the face and a high ejection force on the machine trace.
3. Why did it stick? The die spray missed the lower cavity area. Two of the six spray nozzles were partly blocked. Proof: a spray pattern test on a blotting sheet shows two dry bands.
4. Why were the nozzles blocked? A new die lubricant, introduced on the 2nd, leaves a residue, and nobody had set a nozzle cleaning interval for it. Proof: the purchasing change record and a microscope image of the residue.
5. Why was this change made without a check? Purchasing could swap a consumable to a cheaper grade without a quality review, and the control plan had no spray check. Proof: the change procedure has no step for consumables.

That fifth answer is the root cause the company can act on, because it is a gap in the system, not a person's slip. Now find the escape cause separately. Final inspection measures flatness on five parts at the start of each shift, and at the start of a shift the die is cold and the nozzles are freshly wiped, so the problem only appears after about two hours of running. The sample was taken at the one time of day when the fault could not show.

Bad and good root cause statements

D5 to D7: fix it, prove it, stop it returning

D5, choose actions. Write one for the root cause, one for the escape cause, and one interim action, each with an owner and a date. Interim: clean all six nozzles every four hours and run a spray pattern test at the start of each shift. Permanent for the root cause: return to the approved lubricant and add a quality review to the change procedure for any consumable that touches the product. Permanent for the escape cause: take the flatness sample at two points in the shift, at the start and after two hours of running, and add a 100 percent leak test on cavity B until the capability is proved.

D6, verify. Verification uses the same measure as the complaint, with enough parts to mean something. After the fix, the team measures flatness on 150 consecutive cavity B parts over three shifts and gets a Cpk of 1.64, with the largest reading 0.044 mm. Nothing is closed on a hope: a second check at four weeks confirms the figure and that no leak complaint has come back.

D7, prevent. Look wider than the one part. Update the control plan, the failure mode analysis and the work instruction, then ask which other parts run on dies sprayed from the same system. The same spray fault could appear on any of them.

D8, close. The customer reviews the report, checks the evidence and agrees. The team writes a short lessons-learned entry: change control for consumables, and sampling at the time the fault can show. Recognise the people who did the work; it is part of the method and it makes the next 8D easier.

Common mistakes

• Treating containment as the fix. Containment protects the customer today; it does not stop the next bad part being made.
• No escape cause. If you fix only the cause of the defect, your inspection still cannot catch the next problem.
• A root cause that is a person. Ask why the system allowed it.
• Verification on too few parts, or with a different measure from the complaint.
• Closing the report with no effectiveness check at a later date.
• Changing the document and not the process, so the control plan reads well and nothing happens on the line.

A 12-question checklist before you approve an 8D

• Is the problem stated with is and is-not, and with part and lot numbers?
• Does the containment table add up to the quantity shipped?
• Were all locations checked, including the customer and goods in transit?
• Does the team include someone who can change the process?
• Is every link in the 5-why chain supported by a measurement or an image?
• Is the root cause something that can be changed and checked?
• Is there a separate escape cause with its own action?
• Does each action have an owner and a date?
• Is the verification done on the same measure as the complaint, with a sample size shown?
• Are the control plan, FMEA and work instruction updated?
• Have similar parts and other lines been checked?
• Is there a follow-up date to confirm the result still holds?

How to practise this

Reading a worked example helps, but the skill is writing one yourself under time pressure. The quality engineer job simulation gives you a complaint, measurement data and a customer deadline at a simulated company, and you produce the containment, charts, control plan and failure analysis as tickets that are checked as you hand them in. It sits with the others on the job simulation hub. If your interest is the electrical side, the maintenance electrical engineer simulation includes a root cause report for a plant fault, and the control panel engineer simulation practises the routine test record that sits behind a panel's quality file.

A finished simulation shows on your EDWartens profile under Work experience, marked as a job simulation. The optional certificate has a public page where anyone can verify it. For the wider quality route, read our guide to the electrical QA and QC engineer role and the factory acceptance test checklist for PLC panels. You can see current openings on the jobs board, and the free engineering internships are a guided four-week option with a simulated team.

Last updated 19 October 2026. The case in this post is invented. Follow your own customer's requirements, because they decide the format and the deadline.

Take the free course

Free

Instrumentation · Beginner · Free

Industrial Instrumentation and Process Control

Free

Process control · Intermediate · Free

PID Control for PLC Engineers