Optical Lab KPIs: Benchmarks for Remakes, Turnaround, Yield and Coating

A prescription (Rx) lab should watch six KPIs, each held against two bars: typical performance and world-class. The point is not the score itself but the gap, and the lever that closes it. Pull the raw inputs from your lab management system, then convert them into rates the same way every week.

Every figure shows its basis. A superscript number links to a public source in the Sources list. No public source stating the lab performance ranges below has been verified, so each is marked no cited source: it is kept as originally published and is illustrative, not measured industry data.

At a glance

Remake rate

Remake rate, meaning jobs redone divided by jobs shipped, is the headline quality KPI. World-class labs hold total remakes under 3 percent no cited source, and typical labs sit at 6 to 10 percent no cited source. Troubled labs exceed 12 percent no cited source.

Break it down by cause. Prescription entry and layout errors usually drive 30 to 40 percent of remakes no cited source, coating and scratch defects another 20 to 25 percent no cited source, and breakage during edging the rest.

The highest-leverage fix is a double-check gate on Rx entry, which alone can cut total remakes by 2 to 3 points no cited source.

How to measure

• Divide jobs redone by jobs shipped over the same period.
• Tag every remake with its cause: entry and layout, coating and scratch, or edging breakage.
• Track the entry share on its own.

What moves it

• A double-check gate on Rx entry.
• Coating and handling discipline, for coating and scratch defects.
• Edging practice, for breakage.

Track the share of remakes caused by prescription errors with the eyewear and lens calculators.

Order turnaround

Order turnaround, from receipt to ship, sets your service reputation. Stock single vision should ship same day or next day no cited source.

Surfaced Rx jobs run 2 to 4 days world-class no cited source and 4 to 7 days typical no cited source, and progressives add a day no cited source.

Report turnaround as a median. Add the 95th percentile so stalled jobs show up. Do not use the average, because a few stalled jobs drag it and hide a bimodal queue.

For example, cutting queue by half a day at each of four handoffs recovers two full days.

How to measure

• Clock each job from receipt to ship.
• Split stock, surfaced and progressive jobs before you compute anything.
• Report a median and a high percentile for each group.

What moves it

• Takt discipline, with every station held under the pace of incoming orders.
• Work in process kept from piling up at the generator.
• Shorter queues at each handoff.

Set the pace every station must hold with the Prescription Order Takt calculator.

Surfacing first-pass yield

First-pass yield, lenses good the first time through surfacing, should reach 96 to 98 percent world-class no cited source and sits at 90 to 94 percent typical no cited source. Do not confuse it with blank yield, which counts material scrap before processing.

Low first-pass yield usually traces to generator calibration drift and blocking errors, not bad blanks. The lever is tighter statistical process control (SPC) on power and cylinder, with a control check every 2 hours no cited source.

For example, a lab pulling first-pass yield from 92 to 97 percent removes five first-pass failures per hundred jobs. That is five remakes per hundred jobs if every surfacing failure would have been remade.

Impact resistance is a separate release gate in the U.S. The FDA drop-ball test drops a 5/8-inch steel ball weighing about 0.56 ounce from a height of 50 inches onto the lens¹. A lens fails if it cracks through its entire thickness and across a complete diameter into two or more separate pieces, or if any lens material visible to the naked eye detaches from the ocular surface¹.

Every finished glass prescription lens must be tested individually². For most other impact-resistant lenses, including plastic prescription lenses, the manufacturer must impact-test a statistically significant sample from each production batch².

How to measure

• Divide lenses good the first time through surfacing by lenses surfaced.
• Keep blank yield, which counts material scrap before processing, as a separate number.
• Chart power and cylinder results from each control check.

What moves it

• Generator calibration.
• Blocking accuracy.
• SPC on power and cylinder.

Check surfacing time per job with the Lens Grinding Cycle Time calculator, and confirm the material side separately so you improve the right stage.

Coating chamber utilization

Coating chamber utilization drives return on your most expensive asset. World-class anti-reflective (AR) labs run 78 to 85 percent effective utilization, combining uptime and load factor no cited source.

Typical labs sit at 55 to 65 percent, often because domes ship half empty no cited source. Measure time utilization and load factor separately, because either one can hide behind the other.

For example, a chamber at 90 percent uptime but 60 percent load is really at 54 percent effective, since 0.90 x 0.60 = 0.54.

Likewise, moving from 200 to 230 lenses on 240-capacity domes lifts load factor from 83.3 to 95.8 percent, a gain of 12.5 points.

How to measure

• Log chamber uptime as a share of available hours.
• Log load factor as lenses loaded divided by dome capacity.
• Multiply the two for effective utilization.

What moves it

• Batching by coating recipe to fill domes before a cycle starts.
• Not starting cycles with half-empty domes.

Split utilization into its time and load components with the Coating Chamber Utilization calculator.

Labor productivity

Labor productivity, expressed as good pairs per labor hour, runs 3.0 to 4.0 pairs world-class across a balanced line no cited source and 2.0 to 2.8 typical no cited source.

The lever is line balancing, so no station exceeds takt while others idle.

How to measure

• Divide good pairs shipped by total direct labor hours.
• Measure across the whole line, not one station.

What moves it

• Line balancing against takt.
• Idle time at stations that run ahead of the rest.

Estimate labor per pair for assembly steps with the Frame Assembly Labor calculator.

Coating defect PPM

Coating defect PPM, meaning defective coated lenses per million, should stay under 8,000 world-class no cited source. It often runs 20,000 to 40,000 in average labs no cited source.

Scratches and coating crazing dominate. Handling fixtures and clean-room discipline are the levers.

How to measure

• Divide defective coated lenses by lenses coated, then multiply by one million.
• Tag each defect as a scratch, crazing or another mode.

What moves it

• Handling fixtures.
• Clean-room discipline.

Put a scrap cost on each defect mode, and pick which to attack first, with the Scratch Defect Cost calculator.

Which gap to close first

Sequence improvement by the cost of the gap, not by ease. Each point of remake typically carries the fully loaded cost of a second pair, which is why remakes usually come first.

For example, a lab at 9 percent remakes, 5 day turnaround and 60 percent chamber utilization should attack remakes first. Set a 90 day target of moving remakes from 9 to 6 percent, median turnaround from 5 to 3.5 days and chamber effective utilization from 60 to 72 percent, then take a fresh baseline.

Review the six KPIs weekly on one dashboard, hold each against its typical and world-class bar, and assign one owner and one lever per metric so the numbers actually move.

How to build your own baseline

• Write down each KPI's definition and keep it fixed.
• Pull the raw inputs from your lab management system the same way every week.
• Collect several months of data before you set a target.
• Compare against your own trend first and published ranges second.
• Record what changed in the lab whenever a KPI moves.