Deciding on a corrective action: weighing risk, reach and remedy before you commit

Every decision starts with the user, not the product. Picture the person using the item as intended, and also the foreseeable variations: a tired parent assembling a cot in the evening, a teenager charging a speaker overnight, an older buyer lifting a kettle with wet hands.

Severity and likelihood both count, and they pull in different directions. A fault that rarely bites but bites hard deserves a stronger remedy than a fault that often annoys but never injures. A seller of electric blankets, for instance, treats a controller that overheats as serious even when reports are few, because the user is asleep when the fault develops and cannot react. Naming the harm plainly keeps the rest of the decision honest.

A decision made without a spread map is guesswork. The seller needs the batch or serial range, the sales channels, the countries reached, and whether units went to trade buyers who resold them onward.

Gaps in the map matter as much as the map itself. Gift purchases, second hand resale and bulk sales to workplaces all create holders the seller cannot name from order records. Where the map has holes, the remedy must include a public element, because direct messages alone will miss those holders. Illustration: a maker of desk lamps traces a faulty switch to units sold through two distributors and a marketplace store, and finds the marketplace orders carry buyer messages while the distributor stock went to small shops with no buyer records.

Brainstorming remedies before costing them prevents a familiar trap, which is picking the cheapest option and then defending it. The honest list for most products runs from a posted part, through a workshop repair and a replacement, to a refund and return, plus the option of restricting use, such as disabling a feature, where the product allows it. Each option should be tested against one question: after this remedy, can the fault still hurt someone using the product normally?

Some remedies fail that test on inspection. A warning sticker does not fix a structural weakness. A software limit does not fix a battery that swells. A replacement part that buyers must fit with a tool they may not own is not a remedy for most households. Crossing out the options that fail the test, with a reason written beside each, leaves a shortlist the seller can defend and cost properly.

Speed and reach often decide between two remedies that both remove the risk. A posted part can reach thousands of buyers within days, while a workshop repair programme may take weeks to book everyone in. A refund and return removes the product fastest for buyers who act, but depends on buyers bothering to pack and post, which some will not do for a low value item.

The seller should also weigh the burden placed on the buyer. A remedy that asks a buyer to travel, pay postage upfront, or dismantle a fitted product will see lower uptake, and lower uptake leaves risk in homes. Covering postage, offering collection, or providing a local drop point can matter more than the remedy label. Illustration: a seller of wall mounted drying racks offers free collection of a faulty rack model, because asking buyers to post a bulky rack would have left most of them hanging in bathrooms.

A corrective action decision should not rest with one person working alone. The person who found the fault, the person who understands the product design, and the person who handles buyer contact each see a different part of the picture, and a short joint discussion catches blind spots. The responsible person in the EU and any authority contact should be part of the conversation where the rules or the risk require it, rather than learning of the decision after notices go out.

Once agreed, each strand needs an owner and a date: notices drafted, trade partners briefed, returns process ready, tally started. An ownerless strand drifts, and drift during a correction looks like indifference to buyers and authorities alike. The agreement itself should be minuted in a few lines, with the rejected options and their reasons, so the decision survives staff changes and later scrutiny.

New facts can overturn a sound decision, and the team should watch for them. A further incident, a complaint from a use the team had not considered, or a test result showing the fix does not hold under a common condition all justify reopening. The trigger for reopening should be agreed in advance, so nobody has to argue for a second look while buyers wait.

Reopening is not failure; it is the system working. A seller who upgrades a part replacement to a full return after fresh evidence shows the part fix failing is acting on knowledge, and buyers generally accept a stronger remedy explained plainly. What damages trust is clinging to a remedy the evidence has outgrown, because buyers and authorities eventually notice the gap between the message and the facts.