Jewelry and Watch Manufacturing KPIs: Benchmarks for Metal Yield, Setting and Finishing

Use these benchmarks as reference points for spotting a gap worth investigating, not as targets. In a precious metal shop your own monthly trend per alloy says more than any published range.

Every figure shows its basis. A superscript number links to a public source in the Sources list, and the sentence gives that source's context. A figure marked no cited source is kept as originally published because no public source stating it has been verified, so treat it as illustrative.

At a glance

Precious metal yield

Precious metal yield is the headline KPI because metal is most of your working capital. Typical fine jewelry casting shops run 68 to 75 percent finished-to-issued yield no cited source. World-class operations hold 80 to 85 percent on stable geometries no cited source.

The gap is sprue design, tree density and how tightly polishing loss is controlled. Treat any 3 point drop as a signal that sprue trees grew heavier or scrap segregation slipped no cited source.

For example, on a shop turning 50 kg of gold a year, each point of yield is 500 g, or about 16.08 troy ounces, worth around 38,000 dollars at a spot price of 2,350 dollars an ounce. The value scales directly with the current spot price.

How to measure

• Divide finished metal weight by metal issued over the same period.
• Track it monthly per alloy.

What moves it

• Sprue design and tree density.
• Tight control of polishing loss.
• Scrap segregation, so returned metal is credited to the right alloy.

Track yield per alloy with the Precious Metal Yield calculator.

Casting tree utilization

Casting tree utilization, the product share of the metal on a tree, is the upstream lever behind yield. In a 1995 Santa Fe Symposium paper on casting costs, the main sprue and gates account for half or more of the metal needed, and in most cases that metal can be recycled¹.

Some platinum trees run lower still. On platinum trees with a thick center sprue, a 2007 Platinum Guild International paper reports that many times over 75 percent of the cast is consumed by the sprue and button³.

The best filigree and volume shops reach 74 to 80 percent product by maximizing part count per flask without starving fill no cited source. The lever is sprue diameter discipline and tighter part spacing, not larger flasks.

Returned sprue metal needs care as well. A World Gold Council paper advises using no more than 50 percent scrap in a new melt, and preferably less, especially in investment casting, because the yield of good castings falls as scrap content rises⁴.

For example, moving from 62 to 74 percent utilization on a 300 g flask frees 12 points x 300 g, or about 36 g of metal per pour. Across 20 pours a week that is over 700 g freed from work in process.

How to measure

• Weigh the product and the total tree metal for each flask.
• Report the product share per flask size.

What moves it

• Sprue diameter discipline.
• Tighter part spacing and more parts per flask, without starving fill.
• Clean, known scrap in the melt, kept within the recommended share.

Test tree layouts with the Casting Tree Utilization calculator.

Setting first-pass yield

Setting first-pass yield is the quality KPI that protects both metal and margin. World-class benches hold 97 to 99 percent of stones set correctly with no reset, while typical shops sit at 92 to 96 percent no cited source. Micro pave pulls the average down.

Each reset costs 3 to 6 minutes plus a stone breakage risk of 0.3 to 1 percent no cited source.

For example, a 4 point first-pass gap on a piece with 30 stones means 4 percent of 30 stones, or 1.2 resets per piece on average.

Setting is checked under magnification, and 10x is the U.S. grading reference: the FTC Jewelry Guides allow a diamond to be called flawless only if it shows no flaws under a corrected 10-power magnifier with adequate illumination⁵. Size inspection workload so those checks do not become the bottleneck.

How to measure

• Divide stones set correctly with no reset by stones set.
• Split it by setting style, so micro pave does not hide in the average.

What moves it

• Cut-seat gauging.
• Setter certification, rather than added rework capacity.

Cost setting time, resets included, with the Stone Setting Labor calculator.

Polishing throughput and rework

Polishing throughput and rework rate decide whether finishing is a bottleneck. Benchmark a mirror finish band at 3 to 5 pieces per labor hour and complex openwork at 1.5 to 3 per hour no cited source.

Rework from over-polishing that rounds edges or thins prongs should stay under 2 percent world-class, versus 5 to 8 percent typical no cited source. Flag any bench running more than 15 percent over standard minutes no cited source.

Over-polishing also silently destroys yield, because the metal it removes ends up in sweeps. In a World Gold Council table of typical gold contents, polishing dusts and sweeps hold only 0.5 to 5 percent gold⁴, so that metal returns only through a low-grade stream.

For example, removing 0.05 g per piece across 10,000 pieces is 500 g of gold turned into sweeps.

How to measure

• Pieces finished per labor hour, by finish band.
• Pieces reworked for over-polishing, over pieces polished.
• Actual minutes against standard minutes for each stage.

What moves it

• Stage standards set for each finish band.
• Controlled removal, so edges and prongs keep their form.

Set stage standards with the Polishing Time calculator.

Scrap recovery efficiency

Scrap recovery efficiency is a KPI hiding in plain sight. World-class shops recover 92 to 97 percent of contained metal from clean, segregated scrap no cited source. Poor segregation and mixed alloys drop recovery to 80 to 88 percent and invite refiner disputes no cited source.

Low-grade streams dilute a lot. In the World Gold Council's 2002 table of typical gold contents, bench scrap (lemel) holds 19 to 52 percent gold by weight, and polishing dusts and sweeps only 0.5 to 5 percent⁴.

A 1987 survey of recovery from jewelers' scrap found the same ranges, with lemel usually 28 to 42 percent gold and sweeps usually 1 to 3 percent⁶. Mixing either with clean turnings drags your whole lot assay down.

Verify settlements as well. In a 2011 study, identical 100 g sub-samples of gold grindings came back from refiners at 29.1 to 34.4 percent gold, against the author's own assays of 32.9 to 33.4 percent⁷.

How to measure

• Divide recovered contained weight by estimated contained weight in the lot.
• Keep lemel, sweeps and clean turnings in separate lots.

What moves it

• Segregation by alloy and cleanliness, the single cheapest lever to lift recovered value.
• Split or blind samples to check each refiner's settlement.

Value a lot before it ships with the Scrap Recovery Value calculator.

Plating consistency

Plating consistency and chemistry KPIs keep gold cost and rejects in check. Target plate thickness within plus or minus 10 percent of nominal no cited source.

For example, plus or minus 10 percent around a 2.5 micron nominal is 2.25 to 2.75 micron.

Marketing names set minimums, not nominals. Under the FTC Jewelry Guides, Heavy Gold Electroplate and vermeil need at least 2.5 microns of fine gold equivalent throughout, and Gold Electroplate at least 0.175 microns⁵. For those claims, plate so the thinnest point still meets the minimum.

Bath efficiency varies widely. Generic gold plating-rate tables assume 100 percent cathode efficiency, but a 2001 Santa Fe Symposium paper notes that many gold baths plate at only 50 percent or less, so plating time or current must be extended². Establish each bath's actual efficiency with its supplier and watch for drift from it.

Adhesion and porosity failures should sit under 1 percent for certified goods no cited source. Track drag-out losses too, since 5 to 12 percent of bath gold can leave on parts and rinses if rinse tanks are not recovered no cited source.

How to measure

• Plate thickness at the thinnest measured point, against the spec.
• Actual cathode efficiency, from amp hours and gold deposited.
• Gold recovered from rinse tanks, against gold added to the bath.

What moves it

• Rinse-tank recovery for drag-out.
• Bath chemistry held at the efficiency established with the supplier.
• Rack position and part spacing, which change thickness from piece to piece.

Tie amp hours and efficiency to gold consumed with the Plating Bath Cost calculator.

Traceability and inspection

Traceability and inspection KPIs matter more in luxury than in almost any other sector. For serialized, hallmarked or certified goods, target 100 percent serialization accuracy no cited source and full genealogy from casting lot to shipped unit.

Inspection coverage at 10x to 40x should be defined by AQL, not gut no cited source, with critical defect AQL near 0.65 and major defects around 1.5 for fine jewelry no cited source.

Fineness marks carry legal tolerances. Under the National Stamping Act, as the FTC Jewelry Guides note, gold articles without solder have a permissible fineness tolerance of three parts per thousand, and articles with solder seven parts per thousand⁵.

Watchmaking adds water resistance and timing checks. A COSC-certified chronometer movement is tested individually over 15 days, in five positions and at three temperatures (8, 23 and 38 C), and must hold an average daily rate between minus 4 and plus 6 seconds per day⁸.

COSC certified 2.1 million movements in 2025, and 39 percent of the mechanical watches exported from Switzerland that year were COSC-certified⁹. Individual timing certification applies to those chronometers, not to every watch.

How to measure

• Serial records checked against physical units at audit.
• Genealogy completeness: shipped units traceable to their casting lot.
• Inspection sample sizes and acceptance numbers taken from the AQL tables.

What moves it

• Serial capture at each step rather than reconstruction at shipping.
• Inspection workload sized so full inspection of critical characteristics does not blow throughput.

Repair and return rate

Repair and return rate is the after-sale KPI that reveals build quality. World-class fine jewelry runs a warranty return rate of 1 to 2 percent, typical shops 3 to 5 percent no cited source. Watches vary from 2 to 6 percent depending on complication and water resistance no cited source.

Treat any category running above 5 percent as a design or process defect, not a service cost no cited source. Rhodium re-plating and prong rebuilds dominate jewelry returns, while crown and gasket failures dominate watches.

How to measure

• Returns over units sold, by product category.
• Each return traced back to the bench that built the batch.
• A repair reserve sized against actual returns, not a guess.

What moves it

• Plating thickness, which drives rhodium re-plating returns.
• Gasket torque, which drives crown and gasket failures in watches.

How to build your own baseline

• Write each KPI's definition down, including its period and unit, and keep it fixed.
• Track metal yield monthly per alloy and quality KPIs weekly per bench.
• Compare against your own trend first and published ranges second.
• Record what changed in the process whenever a KPI moves.