After the recall ends: changing the product, the supplier and the checks that missed it

A recall file often names a symptom rather than a cause, and the post recall job starts by separating the two. A handle that snapped is a symptom. The cause might be a thinner wall in one moulding run, a substitute plastic chosen without a strength check, or an assembly step that depended on one worker noticing a misfit. Each of those causes points at a different fix, so guessing wastes the whole exercise.

Proof comes from comparing failed units with sound ones. The team examines broken parts, checks material certificates for the affected run, and replays the assembly sequence with the people who run it. Complaint notes help here too, because the first reports often mention a detail, such as a crack appearing near a screw hole, that narrows the search. Only when the comparison points one way should the team write the cause down as settled.

A good product change attacks the cause directly. If a moulding wall ran thin, the fix is a thicker wall or a different forming method, verified by testing the revised part under the loads it meets in daily use. If a substitute material failed, the fix is a material with a documented strength record for that application, not another untested substitute. If assembly depended on eyesight alone, the fix is a jig or a gauge that makes a misfit impossible to pass on.

Cosmetic changes do not count. A new colour, a stronger looking label or an extra warning leaflet leaves the underlying weakness untouched and may even suggest the problem was solved when it was not. Illustration: a maker of folding step stools replaces a riveted hinge with a bolted hinge plus a locking washer after hinge failures, then load tests the revised stool through repeated folding cycles before any new stock ships.

Sometimes the fault lives in a supplier process the seller never saw. A component maker might have changed a sub supplier, skipped a curing step, or shipped mixed batches under one code. In that case the seller needs more than an apology: a written account of what changed, evidence that the new process is controlled, and a trial order inspected on arrival before normal ordering resumes.

Not every fault justifies ending a supplier relationship. A supplier who discloses the lapse promptly, shows corrected controls and accepts incoming inspection may be safer than an unknown replacement. The deciding test is evidence of control going forward, not the warmth of the conversation. Where a supplier cannot show that evidence, the seller should qualify an alternative and keep the old supplier out of safety critical parts until proof exists.

Every recalled fault slipped past some check, and the post recall review should name that check honestly. Perhaps incoming inspection sampled too few units, or a finished product check looked at appearance while the fault hid inside a joint. Perhaps no check existed for the property that failed, because nobody had listed it as a risk for that product type.

The revised check should be specific enough for a new inspector to run it without asking. It states the property, the method, the sample, and the action when a unit fails, and it records results per batch so a drift shows up before buyers notice. Illustration: a seller of ceramic mugs adds a handle pull test to every batch after handles parted from mugs, logging the result against the batch code so a weak batch can be held before dispatch.

A recall exposes gaps in who knew what. Perhaps complaints sat in a sales inbox nobody reviewed for safety, or the person who could halt shipments was on leave with no deputy named. Fixing that means naming a safety contact and a deputy, giving them authority to stop dispatch, and routing complaints through one register where safety signals are flagged rather than filed as general feedback.

Documents change too. The risk assessment gets the newly found hazard added with the revised controls, the technical file gains the test evidence for the redesign, and the supplier file gains the new inspection terms. These updates matter because the next product version inherits them; a lesson stored only in memory will be lost when staff move on, while a lesson stored in the file travels with the product.

Proof accumulates quietly after the change. Complaints about the old fault should fall away, incoming inspection should show steady results, and any fresh failure should look different from the recalled one. The team should set a review date, perhaps after a full selling season, and compare the new record with the old one rather than declaring victory on launch day.

That review closes the loop properly. It records the comparison, notes any remaining weakness, and either retires the issue or sets a further change. A seller who can show that sequence, fault, cause, change, evidence, review, holds a stronger position with any authority or trade partner who later asks what changed, and more importantly holds a product that buyers can trust again.