A London boutique bought 250 pairs of silver moissanite studs at a trade show. Within 90 days, 42 pairs came back — loose stones, mismatched pairs, dull plating. That is a 17% return rate, which on a 30% margin wipes out the whole order's profit and then some. Most of those returns were catchable at the factory before shipping. Below is the QC checklist we run on every parcel, and the version you can apply the moment your own incoming goods arrive.
Where the returns actually come from
Internal defect data across our last 1,000 wholesale parcels puts the cause distribution like this:
| Defect cause | Share of returns | Catchable at factory? |
|---|---|---|
| Loose stones / prong failure | 38% | Yes — pull test |
| Plating wear / yellowing | 22% | Yes — thickness gauge |
| Mismatched pairs / sizes | 14% | Yes — carton audit |
| Surface scratches / polish marks | 12% | Yes — 10x loupe check |
| Engraving / logo errors | 8% | Yes — pre-pack proof |
| Other (alloy, stones, size) | 6% | Mostly yes |
Five inspection points cover 94% of what your customers will complain about.
The six-point factory QC checklist
1. Incoming alloy XRF
Before casting, every alloy lot is XRF-scanned. Sterling silver must read 92.5% Ag; 14k gold must read 58.3% Au. We strip and re-melt any lot outside +/- 0.5%. Ask your supplier for the strip — a missing XRF log is a missing QC step.
2. Casting and tree inspection
Greeners, porosity, shrinkage and flash are rejected at the tumbling stage. A porous casting may look fine on the surface but will crack at a prong within months. We reject 1.5–2% of castings here; a factory that rejects nothing is not inspecting.
3. Prong pull test
Every pavé and prong-set stone receives a 1.5 kg pull on a calibrated gauge. Industry norm is a visual close with no pull. Our rejection rate on this step is under 0.5%, but when it fails, the stone never leaves the bench.
4. Plating thickness
Rhodium is measured on a coulometric gauge each shift. We plate to 0.5 µm; the industry commonly runs 0.3 µm. The extra 0.2 µm is the difference between white finish lasting 6 months and 14 months on a customer's finger.
5. Final visual and dimensional check
100% visual under 10x loupe: scratches, symmetry, engraving legibility, prong smoothness. A 10% random sample is caliper-checked against the approved sample — shank thickness, inner diameter, stone seat depth. Tolerance is +/- 0.10 mm.
6. Carton audit before sealing
Pieces are counted against the packing list, pairs matched, sizes sorted. Carton weight is compared to expected weight. This single step kills the "mismatched pair" category of returns almost entirely.
Run your own incoming inspection
When your parcel lands, spend 20 minutes on a 10% random sample before you sign off: pull-test a few pavé prongs with tweezers, check plating under a 10x loupe, caliper one shank per SKU, and confirm pair matching. If more than 2% of the sample fails, file an RMA within the 30-day window. We accept defect returns in that window with a re-make or credit — no debate.
Tailor the depth to your order
Small opening orders (under 200 pcs): run points 3, 5 and 6 yourself on arrival. These three catch the defects customers actually notice.
Growing boutiques (200–1,500 pcs): request the XRF strip and plating log with every shipment. These two documents prove the factory ran the process, not just the visual check.
High-volume brands (1,500+ pcs): commission a third-party pre-shipment inspection (SGS/ITS, AQL 2.5) on top of our internal QC. The cost is USD 200–300 per lot, and it insures a five-figure shipment against a bad batch.
Frequently Asked Questions
What is a good defect rate for wholesale jewelry?
Below 2% on arrival is healthy. Above 5% means the factory is skipping gates; above 10% means you are buying from a trading shell. Reject the supplier.
How do I pull-test a prong without a gauge?
Use a pair of fine tweezers and steady pressure. A properly set pavé stone will not shift; a loose stone moves under gentle force. It is a screen, not a measurement.
What does AQL 2.5 mean for jewelry?
AQL 2.5 is the standard textile/general-inspection sampling level. For jewelry, it means the third-party inspector samples a fixed number of pieces and accepts the lot only if critical defects stay below the accepted limit.
How long is a defect RMA window?
We accept written defect claims within 30 days of delivery, with photo evidence. Stone loss within 90 days on properly worn pieces is covered under the prong-set warranty.
Should I inspect every piece on arrival?
No. A 10% random sample per SKU, focused on the six points above, catches systemic defects. Reserve 100% inspection for engraved or logo-critical orders.
Key takeaways
- 94% of customer returns trace to six catchable factory points.
- Demand XRF strips, 0.5 µm plating logs and prong pull records with every lot.
- Run a 10% incoming sample check before signing off a new supplier.
- Above 1,500 pieces per lot, add a third-party AQL 2.5 inspection.
Next step: ask for the last 90 days of our QC log with your quote. Sample parcels ship in 5 business days; MOQ 20 pcs per SKU.
Related: wholesale factory tour, OEM manufacturing, moissanite, and all QC guides.
For third-party reference, the Wikipedia article on moissanite details moissanite’s discovery, optical properties and use in jewellery.
