The 8D problem solving method has been around since the late 1980s, and most quality engineers can recite the eight disciplines from memory. Team, problem, containment, root cause, corrective action, implementation, prevention, closure. Knowing the list is easy. Running an 8D that a customer accepts on the first submission, and that actually stops the defect from coming back, is a different skill.
This guide walks through each of the 8 disciplines the way they play out on a real shop floor. For every step you'll get what the step is for, what "good" looks like, and the mistake that most often gets a report bounced back. We'll follow one example the whole way through so you can see how the disciplines connect.
What is 8D problem solving?
8D (Eight Disciplines) is a structured, team-based method for solving a quality problem and preventing it from recurring. It was popularized by Ford in its Team Oriented Problem Solving manual, and it's now the default format for customer complaints across automotive, aerospace, medical devices and general industry. If you supply an OEM, there's a good chance your customer requires an 8D for every complaint above a certain severity.
The method is built on a simple idea: protect the customer first, then find the real cause, then fix the system so the problem can't happen again. Every discipline exists to stop you from skipping one of those three.
When should you use an 8D?
Not every problem deserves an 8D. A full 8D ties up a cross-functional team for days or weeks, so save it for problems that are:
- Reported by a customer, or serious enough that a customer would care
- Recurring, or likely to recur
- Unclear in cause. If you already know exactly what happened and the fix is obvious, a quick corrective action record is enough
- Big enough in cost, safety or volume to justify the team's time
For small, local issues a 5-Why on a whiteboard is often the right tool. For an issue your customer is tracking with a deadline, use the 8D.
The running example
Throughout this guide we'll use one problem. A Tier 2 supplier molds a glass-filled nylon (PA66) connector housing. The customer reports that 14 housings out of a 2,000-piece lot cracked at the screw boss during assembly. The parts passed outgoing inspection.
D0: Prepare (optional, but worth doing)
Some formats, notably Ford's Global 8D, add a D0 step before the team is even formed. D0 asks two questions. Does this problem need an 8D at all? And what emergency response is needed right now to protect the customer?
Even if your customer doesn't require D0, the thinking is useful. In our example, the emergency response is a phone call to the customer within hours, an offer to sort their stock, and a hold on everything in the warehouse from the same material lot.
D1: Establish the team
Purpose: put the right people on the problem, with the authority to act.
A good D1 has three roles filled:
- Champion. A manager who can approve spending, free up people and remove roadblocks. They don't do the analysis.
- Team leader. Usually the quality engineer. Owns the report and the timeline.
- Members. People who actually know the process: the process or molding engineer, a production supervisor, maintenance, and someone from the operator team who runs the job.
Common mistake: a team made only of quality people. Quality engineers are good at structure, but the person who knows that the dryer alarm has been "acting up for weeks" is usually on the floor. If your team has no one who touches the process every day, your D4 will be weaker for it.
Example D1: Champion: Plant Manager. Team leader: Quality Engineer. Members: Molding Process Engineer, Second-Shift Production Supervisor, Maintenance Technician, Molding Operator.
D2: Describe the problem
Purpose: define the problem precisely enough that you'll know when it's solved.
This is the step most teams rush, and it causes more weak 8Ds than any other. A strong D2 answers the 5W2H questions (what, where, when, who, why it matters, how, how many) and ideally includes an Is / Is Not comparison.
Weak D2: "Customer found cracked housings."
Strong D2: "14 of 2,000 PA66-GF30 connector housings (P/N 7731-B, lot L2611-04, molded 3 November, second shift, press 6) cracked at the M3 screw boss during customer assembly on 5 November. Cracks are radial, originate at the boss root, and appear only after screw insertion. Defect rate 0.7%. Specification: no cracks after insertion at 0.6 Nm."
The Is / Is Not analysis narrows the search before anyone starts guessing:
| Is | Is Not | |
|---|---|---|
| What | Cracked screw boss | Short shots, flash, dimensional issues |
| Where | Press 6 | Presses 4 and 5 (same tool family) |
| When | Lot L2611-04, second shift | First shift, previous lots |
| How many | 14 of 2,000 | Zero in prior 20 lots |
Common mistake: writing a cause into the problem statement. "Housings cracked because of a bad material lot" is not a problem description. It's a guess, and it biases everything after it.
D3: Interim containment actions
Purpose: protect the customer while you look for the root cause.
Containment is temporary by design. It should be fast, it should cover every location where suspect product exists, and it should be verified.
A complete D3 usually covers:
- Stock at the customer, in transit, in your warehouse and in WIP
- How suspect parts are sorted (method, gauge, criteria)
- How good parts are identified (a mark, a label, a certificate)
- Proof the containment works. If you're sorting, how do you know the sort catches the defect?
Example D3: All lot L2611-04 stock quarantined: 1,400 pcs at customer, 3,000 pcs in warehouse. 100% screw insertion test at 0.8 Nm (above the customer's 0.6 Nm) on all suspect parts. Certified parts marked with a green dot on the label. Sort verified by seeding five known-cracked parts; all five detected.
Common mistake: forgetting the pipeline. Parts in transit and parts sitting in a customer's line-side rack are the ones that escape containment most often.
D4: Root cause analysis and verification
Purpose: find the true cause, prove it, and explain why your controls didn't catch it.
D4 is the heart of the 8D, and it's where customers push back hardest. A strong D4 has two separate root causes:
- Occurrence root cause. Why did the defect happen?
- Escape root cause (also called non-detection). Why didn't you catch it before it left the building?
Tools like the Ishikawa (fishbone) diagram help the team brainstorm possible causes across man, machine, method, material, measurement and environment. A 5-Why then drills down the most likely branch. Most importantly, the cause has to be verified: you should be able to switch the defect on and off by switching the cause on and off.
Example D4:
- Occurrence: The material dryer on press 6 ran with a dew point of -10°C instead of the required -40°C after a desiccant regeneration fault. The dryer alarm had been bypassed by maintenance two weeks earlier during a sensor issue and never restored. PA66 molded with excess moisture degrades and becomes brittle at the boss.
- Verification: Parts molded from material dried to the correct dew point passed 50 of 50 insertion tests. Parts molded from deliberately under-dried material reproduced the cracking in 6 of 50 parts.
- Escape: The outgoing check is visual only. Cracks appear only after screw insertion, so a visual check cannot detect them, and there was no moisture check on material at the press.
We cover this step in depth in our guide on how to write a D4 root cause.
Common mistake: stopping at "operator error" or "material issue." If your root cause can't be verified, or if it could be fixed with "retrain the operator," keep digging.
D5: Choose and verify permanent corrective actions
Purpose: select fixes that remove each root cause, and prove they work before full rollout.
Every root cause in D4 needs at least one action in D5. That includes the escape cause. For each action, record the owner, the due date and how you'll verify it.
Example D5:
- Restore the dryer dew point alarm and add an interlock that stops the material loader if dew point exceeds -30°C. (Occurrence)
- Add a moisture analyzer check at press start-up and every four hours, with a limit of 0.2%. (Occurrence and escape)
- Add a sample insertion test of 5 pcs per hour at 0.8 Nm to the control plan for this part family. (Escape)
Verification: a trial run of 2,000 pieces with the interlock active and the new checks in place produced zero cracks.
Common mistake: choosing actions that depend on people remembering. An interlock beats a reminder every time.
D6: Implement and validate
Purpose: roll out the permanent actions, remove the containment, and prove the problem is gone over time.
D5 proves the action works in a trial. D6 proves it works in normal production. Define a validation period and a measurable success criterion before you start.
Example D6: Interlock and moisture checks live on 12 November. Validation period: six weeks of production. Result: zero cracked bosses at the customer and zero failures in the hourly insertion test. Containment sort removed on 24 December after customer approval.
Common mistake: declaring victory after a few days. Intermittent problems need a validation window long enough to cover the conditions that caused them, like a full material lot cycle or a seasonal humidity swing.
D7: Prevent recurrence
Purpose: fix the system that allowed the problem, and spread the lesson to similar products and processes.
D7 is where good 8Ds become great ones. The question isn't "will this exact defect come back?" but "where else could the same weakness bite us?"
Typical D7 actions:
- Update the PFMEA for the drying process with the new failure mode and controls
- Update the control plan and work instructions
- Change the maintenance procedure so alarms can't be bypassed without a documented, time-limited deviation
- Read across to every press and dryer in the plant, not just press 6
You can use our free FMEA template and control plan template to make the document updates quickly.
Common mistake: leaving D7 blank or writing "lessons learned shared with team." Name the documents you changed and the other lines you checked.
D8: Closure and team recognition
Purpose: confirm everything is complete, get sign-off and recognize the team.
Closure isn't a formality. Before signing, the champion should confirm that the root cause was verified, every action is complete, validation data is attached and the customer has accepted the report. Then thank the team properly. People remember whether their effort was noticed.
Example D8: All actions closed and verified. Customer accepted the 8D on 3 January. Report signed by the Champion and the Quality Manager. Team recognized at the monthly plant meeting.
Typical 8D timelines
Customers vary, but many OEMs expect something like this:
| Milestone | Typical expectation |
|---|---|
| Acknowledgement and D3 containment | 24 hours |
| D4 root cause identified | 5 to 10 working days |
| D5 and D6 actions defined and started | 30 days |
| Full closure | 60 to 90 days, depending on validation |
Always check your customer-specific requirements. Missing a containment deadline usually hurts your supplier rating more than a slow root cause.
A quick summary of the 8 disciplines
| Discipline | Key question | Output |
|---|---|---|
| D1 Team | Who can solve this? | Named champion, leader and members |
| D2 Problem | What exactly is wrong? | 5W2H statement, Is / Is Not |
| D3 Containment | How do we protect the customer now? | Sorted, identified, verified stock |
| D4 Root cause | Why did it happen and escape? | Verified occurrence and escape causes |
| D5 Corrective action | What removes each cause? | Actions with owners and verification |
| D6 Implementation | Does it work in production? | Validation data, containment removed |
| D7 Prevention | Where else could this happen? | Updated FMEA, control plan, read-across |
| D8 Closure | Is everything done? | Sign-off and recognition |
Getting started
If you want a structure to work from, our free 8D report template walks you through each discipline and exports a clean document. To see a full report written out end to end, read our complete 8D report example. And if your team writes a lot of 8Ds, Vantage 8D can turn your rough notes for each discipline into a polished, customer-ready report and track the resulting actions until they're closed.