Accept, Rework or Reject: Handling a Mixed Jewelry Batch

The inspection is finished and the result is untidy. Most of the delivery is clean. A group of pieces has a fault you can live with. A smaller group has one you cannot. Rejecting a jewelry batch outright would be simple and would also strand stock you have already paid freight and duty on, while accepting the whole thing pushes the problem onto customers who will report it back to you one review at a time.

Below we set out the outcomes actually available to you, the three questions that separate them, and how to record the decision so it holds when the next delivery arrives.

Sort first, decide second

No decision is possible while the pieces are mixed. Separate the delivery into physical groups on the bench: passes, pieces with one named fault, pieces with a different named fault, and pieces with more than one. Count each group and label it.

Sorting frequently changes the answer on its own. A fault that felt like a batch failure often turns out to sit on a handful of pieces from one carton, which points at a single session at the bench rather than at the model. Naming each fault precisely matters here, and the vocabulary for that is already set out in the jewelry defect atlas.

The outcomes you can actually choose from

Buyers tend to think in two options when there are more.

• Accept as delivered, with the fault noted for the next run.
• Accept the passing pieces and hold the rest, so sellable stock reaches the listing without waiting.
• Rework locally, using a bench near you for faults that need a small correction.
• Return the failing pieces for rework and re-delivery, leaving the rest with you.
• Return everything and treat the run as failed.

The last option is rare and expensive in time. It is reserved for faults that run through the whole population, not for a bad group inside an otherwise sound delivery.

Question one: would the customer ever see it?

Judge each fault from the position a wearer occupies, not from under your magnifier. A mark on the inside of a shank, a tool trace on the back of a bezel and a slightly heavier seam inside a clasp are working surfaces. A mark on the front face of a pendant is not.

Faults that a customer would never notice belong in your defect log rather than in a rejection. Escalating them costs goodwill you will want later for a fault that genuinely matters, and it trains a supplier to treat all your feedback as background noise.

Question two: is it correctable without remaking the piece?

Some faults are surface work. Repolishing, replating, dressing an edge, closing and soldering a ring, resetting a stone that moved. Those pieces are candidates for rework wherever a competent bench is available.

Other faults are structural. Wrong alloy, wrong dimensions, a misspelled name, thin walls, a crack running from a stress point. Those pieces cannot be corrected; they can only be replaced or melted back. What happens to them at the workshop is described in how a faulty piece is handled at the bench. Sizing faults deserve their own treatment, since not every product survives a size change, as the guide to rings that arrive in the wrong size explains.

Question three: where is the piece, and what does moving it cost you?

Geography drives more of this decision than buyers expect. Pieces already cleared into your market carry freight and duty that a return journey repeats in both directions, and a second import when they come back. Pieces still with the supplier carry none of that.

That asymmetry has a practical consequence. Faults you can only fix at the origin are worth catching before dispatch, which is the argument for a pre-shipment check rather than a purely incoming one. Once the goods have landed, local rework is often the faster route even for work the supplier would have done without argument.

Rework at your end or at theirs

Local rework buys speed and keeps stock in the country you sell from. It costs you the ability to hold the supplier to the original standard, because once another bench has touched the piece, the fault is no longer cleanly theirs.

Supplier rework preserves that line and usually produces a better technical result, since the workshop has the mould, the finish process and the reference piece. Agree in writing which route applies before you start moving anything, and agree how the reworked pieces will be identified when they come back so they can be checked as a group.

When rejecting the whole batch is right

Three situations justify it. The wrong alloy across the population. A dimensional error that runs through every piece, which points at the mould or the drawing revision. A fault that makes the product unsafe to wear.

In each case the pieces are not repairable at scale and partial acceptance would leave you selling two versions of the same listing. Say so plainly, name the population, and state which revision of the specification the pieces failed against.

Writing the decision so it holds

Record the counts per group, the fault named for each, the outcome chosen, and what you expect from the supplier as a result. Send it as one message rather than as a series of complaints, and keep the tone factual; you are describing a batch, not a relationship. Our contact page is where a workshop conversation about a specific delivery starts.

Use a simple order of operations the next time a delivery splits. Sort, then release everything that passes so your listings keep running. Then take the failing group and ask whether the fault is visible, correctable, and cheaper to fix here or there — in that order. Only when the same fault appears across the whole population does the question stop being about pieces and become a question about the run itself.