8D D5: choose and verify the permanent corrective action before rollout

D5 is the decision step. D4 named the causes; D5 says which permanent action will remove them and shows, before rollout, that the action actually does. A reviewer reads D5 for three things: does each action match a verified root cause, why this action and not another, and what proof exists that it works without creating a new problem. Formats differ by customer, so check their 8D template and supplier manual.

This page is about D5 itself. For how D5 differs from D3 containment, see corrective action vs containment; for how 8D relates to a CAPA system, see CAPA vs 8D.

One action per root cause, on both sides

A complete D4 usually has two causes: why the defect was made (occurrence) and why it wasn’t caught (escape). D5 needs an action for each. Fixing only occurrence leaves the same detection gap for the next cause; fixing only detection leaves the process making defects. Map them in a small table: root cause, chosen action, owner, verification. If D4 is weak, D5 will be too, so check it first with weak D4 root cause examples.

D5 checklist: what the section should contain

• Link to D4: each action names the occurrence or escape cause it removes.
• Options considered: two or three candidate actions with a short reason for the choice, such as effect on the cause, feasibility, lead time and cost of change.
• Error-proofing first: a design, tooling, parameter or poka-yoke change ranks above a procedure, and a procedure above a reminder.
• Verification before rollout: trial run, pilot lot, capability study, gauge check or validation test, with the acceptance criterion stated in advance.
• Side-effect review: what else the change could affect, such as other characteristics, cycle time, other part numbers or customer-approved processes, and how that was checked.
• Change approvals: whether the change needs customer notification or approval under your change process, and its status.
• Owner and planned implementation date carried into D6.

Verifying before implementation

Verification in D5 answers “does this action remove the cause?” on a small scale, before the full rollout. Typical evidence:

• A trial on the affected machine or line with the defect condition reproduced, then shown absent after the change.
• A pilot lot, fully inspected, with the result recorded.
• A capability or measurement system check for the characteristic the action targets.
• For a detection action, a challenge test with known bad parts to show the check catches them.

Set the pass criterion before the trial. “Trial OK” without a criterion or data is one of the first things a reviewer queries.

Why “retrain operator” and “add inspection” are weak

Retraining assumes the cause is a person, which a reviewer rarely accepts without proof that the method, tooling or instruction was sound. See why operator error is not a root cause. Adding an inspection is often containment wearing a D5 label: it catches defects but doesn’t stop them being made, and manual checks are known to miss. Both can support a stronger action; on their own they tend to draw a Conditional or Reject response.

Weak vs strong D5: illustrative examples from different sectors

These examples are fictional and only show the structure.

• Automotive stamping: weak: “Operators retrained, full visual check added.” Strong: “Occurrence: die guide pin replaced with hardened design; escape: in-die sensor stops press on misfeed. Options: sensor vs end-of-line camera; sensor chosen as it stops at source. Pilot of 2,000 strokes with zero misfeeds against a criterion of zero; challenge test with 10 seeded misfeeds, all stopped.”
• Electronics: weak: “AOI threshold tightened.” Strong: “Occurrence: stencil aperture reduced for the package; escape: AOI library updated and challenged with known bad boards. Pilot of three panels, X-ray sample checked against the agreed void limit.”
• Medical device molding: weak: “Process adjusted.” Strong: “Hold pressure window narrowed after a designed trial; change assessed under change control; dimensional study run before release to production.”
• Food packaging: weak: “Line staff reminded to check seals.” Strong: “Sealing jaw temperature control replaced; seal strength tested on a pilot run against the spec; checkweigher reject function tested with seeded faults.”

How D5 hands off to D6

D5 proves the action works on a trial. D6 implements it fully and shows it keeps working in normal production over an agreed period, then ends the interim containment. Write D5 so D6 can pick it up: name the action, the implementation date, the measure you will watch and the period. Details are in D6 verification of effectiveness. Updating the FMEA, control plan and similar products comes after, in D7 prevent recurrence. The free printable 8D / D4 checklist (PDF) covers D1–D8.

FAQ

Is D5 the same as implementing the action?

In many 8D formats D5 is choosing and verifying the permanent corrective action, and D6 is implementing it and confirming it works over time. Check your customer’s format, as some combine the two.

Can D5 have more than one action?

Yes. Usually it should: one action for the occurrence root cause and one for the escape root cause, each with its own verification.

What does 8D Gate check?

It gives a pre-send readiness check on the 8D text: an Accept / Conditional / Reject advisory with missing evidence and questions. It isn’t a customer decision. The review tool is v0.1: it reviews English 8D reports, and the interface and report are in English. Templates are also available in German, Czech and Polish.

Free pre-send readiness check

Before the 8D goes out, run the text through the free supplier 8D readiness check.

Get the free 8D / D4 checklist

Keep the one-page version ready for the next complaint. It covers D1–D8 and goes deeper on containment, D4 root cause, and verification.