LED Luminaire Manufacturing KPIs and Benchmark Ranges
The KPIs that run an LED fixture plant: how to measure each, the levers that move it, and which benchmark figures have a cited public source.
Use these benchmarks as reference points for spotting a gap, not as targets. Your own trend by line, platform and build month says more about your plant than any single published figure, so let the ranges below point you to the loss worth investigating first.
Every figure shows its basis. A superscript number links to a public source in the Sources list, and the sentence says whether that source measures lighting plants or manufacturing in general. A figure marked no cited source is kept exactly as originally published because no public source stating it was found, so treat it as illustrative.
At a glance
LED luminaire manufacturing KPI benchmarks at a glance
Words in parentheses give each figure's original framing or the source's scope. Not stated means the original guide gave no figure for that column. Burn-in fallout of 1 to 2 percent signals a supplier or thermal problem no cited source. The cited OEE figures describe manufacturing in general, not SMT lines specifically1.
Method: How the figures in this guide are sourced
A figure with a superscript number is stated by the cited source, listed under Sources with its publisher, location in the document and access date, and the sentence gives its scope. The cited OEE, warranty and queueing figures are cross-industry, not specific to lighting. A figure marked no cited source is kept exactly as originally published because no public source stating it was found. It is illustrative, not measured industry data. Worked examples show arithmetic or a scenario only and are not benchmarks.
First-pass yield at final assembly
Lead with first-pass yield at final assembly, the share of fixtures passing all in-line and end-of-line tests without rework. Typical plants sit at 92 to 96 percent no cited source, and world-class lighting lines hold 98.5 percent or better no cited source.
The dominant loss modes are cold solder joints, driver connector faults and gasket misplacement. The fastest levers are poka-yoke fixturing on connector orientation and in-line in-circuit test (ICT) before final close-up, which together commonly move a 94 percent line above 97 within a quarter no cited source.
How to measure
Divide good units out by units started.
Count a unit that touched a rework bench as a fail, even if it later passes.
Sort failures by loss mode so the biggest one is always visible.
What moves it
Poka-yoke fixturing on connector orientation.
In-line ICT before final close-up.
Targeted fixes for cold solder joints, driver connector faults and gasket misplacement.
Burn-in fallout rate is the fraction of soaked units that fail during or after the burn-in soak. Best-in-class driver-level fallout runs under 0.3 percent no cited source, and 1 to 2 percent signals a supplier or thermal problem worth a containment no cited source.
Track infant mortality in the first 2 hours no cited source separately from late failures, since the two point to different root causes. Infant mortality usually means solder or handling damage; late failure usually means marginal components or overtemperature.
How to measure
Divide units failing during or after the soak by units soaked.
Split failures into infant mortality and late failures.
Report fallout at driver level as well as fixture level.
What moves it
Lower drive current to pull driver Tc back under its rated point.
Tighter incoming electrical inspection.
Containment and supplier action as soon as fallout jumps.
Field warranty return rate is the KPI customers actually judge you on. Quality LED fixture programs target under 1 percent cumulative over a 5 year warranty no cited source; 2 to 4 percent is common on price-driven products and quietly destroys margin no cited source.
The only broad published comparison is a cost measure. Across roughly 1,400 U.S.-based manufacturers filing with the SEC, warranty claims averaged 1.30 percent of product sales in 20254. That is a dollar figure across all industries, not a unit return rate for lighting.
Driver quality and thermal headroom are the strongest levers. In a survey of members of the LED Systems Reliability Consortium, run with the U.S. Department of Energy, power and driver components were the most commonly observed LED luminaire failure, ahead of LED and moisture failures5. Failed electrolytic or film capacitors are a leading cause of driver failure, and derating those capacitors extends life5.
That makes driver thermal margin the single highest-return design decision: specify 15 to 20 C of Tc margin at worst-case ambient no cited source.
How to measure
Count returns per thousand shipped, by production month.
Normalize by shipment cohort so each build month is judged on its own units.
Tie any spike to a specific build window and lot.
What moves it
Driver quality and supplier selection.
Driver case-temperature (Tc) margin at worst-case ambient.
Overall equipment effectiveness (OEE) on the surface-mount technology (SMT) line, and an OEE equivalent on final assembly, frame your throughput health. World-class OEE is 85 percent or higher2, 3. That benchmark traces to Seiichi Nakajima's total productive maintenance work and applies to discrete manufacturing in general, not SMT specifically.
Measured averages sit well below it. A Swedish study of manufacturing machines found OEE averaging 51.5 percent, and published averages range from 40 to 65 percent1. Vorne Industries, which sells OEE monitoring systems, reports that most manufacturers score closer to 60 percent3.
The high end of measured results comes with a caveat. A study of 23 Swedish companies found a median OEE of 70 percent and a mean of 65 percent, but its author cautions that recorded performance was likely overstated6.
On SMT lines the loss is usually availability, driven by feeder changeovers and reflow profile changes. Research on printed circuit board assembly by Trovinger and Bohn found that as much as 50 percent of effective capacity can be lost to setups7. In the same study, combining single-minute exchange of die (SMED) with IT-based setup tools cut key setup times by more than 80 percent7.
For example, attacking changeover with SMED, cutting a 40 minute feeder swap to 12 minutes, can lift a 68 percent OEE line into the high 70s without any new equipment, depending on how often the line changes over.
On final assembly, track takt attainment, the percent of units leaving within takt; world class is above 95 percent no cited source.
At plant level, U.S. electric lighting equipment makers ran at 88.5 percent of full production capability in the fourth quarter of 2025, against 74.6 percent for all manufacturing8. That comes from the Census Bureau's quarterly capacity survey and measures capacity use, not OEE.
How to measure
OEE: availability times performance times quality, per SMT line3.
Log every feeder changeover and reflow profile change as availability loss.
Takt attainment: units leaving final assembly within takt divided by all units.
Test bottleneck utilization deserves its own KPI because photometric and burn-in assets are expensive and easy to overload. Keep goniophotometer and integrating-sphere utilization in a 70 to 85 percent band no cited source.
Above 90 percent, queue time explodes and test becomes the constraint that starves shipping9. Queueing theory explains the shape rather than a threshold: as arrivals approach a station's service rate, work in process and time in system increase slowly at first and then rapidly9.
For example, in the simplest single-server queue the average wait is u / (1 - u) service times at utilization u: about 2.3 at 70 percent, 5.7 at 85 percent, 9 at 90 percent and 19 at 95 percent.
How to measure
Divide scheduled test hours by available bench hours each week.
Track the goniophotometer and the integrating sphere separately.
Watch the queue in front of each asset, not just its utilization.
What moves it
Shifting quick-check volume from the goniophotometer to the integrating sphere.
Tightening the sampling plan where AQL allows.
Batching same-SKU tests to cut setup between recipes.
SKU complexity is a silent KPI that drags every other number down. Track active SKUs per platform and the ratio of SKUs generating 80 percent of revenue, where healthy lines find 20 to 30 percent of SKUs drive 80 percent of volume no cited source.
A long tail of low-runners inflates changeovers, test recipes and inventory. Measure changeover frequency and average run length alongside the SKU count.
The lever is disciplined rationalization plus attribute pooling. One programmable driver spanning three lumen packages, for example, shrinks the qualification matrix and lifts run length, often improving OEE and yield as a side effect.
How to measure
Count active SKUs per platform.
Rank SKUs by volume and revenue to separate the short head from the long tail.
Track changeover frequency and average run length alongside.
What moves it
Disciplined SKU rationalization.
Attribute pooling, such as one programmable driver across several lumen packages.
A smaller qualification matrix and fewer test recipes per platform.
Close the loop with cost of quality as a percent of revenue, the umbrella metric that ties yield, fallout and returns to money. World-class lighting operations hold total cost of quality, meaning scrap plus rework plus warranty, under 3 percent of revenue no cited source; 5 to 8 percent is common and is where most margin leaks hide no cited source.
Only the warranty part has a published cross-industry average: claims came to 1.30 percent of product sales for U.S. manufacturers in 20254.
Measure it monthly and split it into prevention, appraisal and failure buckets. A plant spending heavily on appraisal, such as burning in every unit, but light on prevention is paying to inspect in quality it should have built in. Shifting spend toward prevention is the durable path to every benchmark above.
How to measure
Add scrap, rework and warranty cost each month and divide by revenue.
Split the total into prevention, appraisal and failure buckets.
What moves it
Shifting spend from appraisal to prevention.
Every yield, fallout and return lever in the sections above.
Anatomization of the overall equipment effectiveness (OEE) for various machines in a tool and die shop. Journal of Intelligent Manufacturing and Special Equipment, Vol. 3 No. 1, pp. 97-105 (Emerald, CC BY 4.0), DOI 10.1108/JIMSE-01-2022-0004. published 2022. p. 101 (PDF p. 5): sentence "Seiichi Nakajima included the above four “world-class” numbers." and Table 1 "The world-class OEE factor"; p. 103 (PDF p. 7): "worldwide threshold of >85% set by the Japan Institute of Plant Maintenance (JIPM)". Accessed September 30, 2026.
World-Class OEE: Set Targets To Drive Improvement. Vorne Industries (OEE.com). Sections 'What Is World-Class OEE?', 'TPM and World-Class OEE', 'Some Perspective on World-Class OEE'. Accessed September 30, 2026.
23rd Annual Product Warranty Report. Warranty Week (trade newsletter; data from U.S.-based manufacturers' SEC 10-K filings). published April 16, 2026. Headline summary and section 'Figure 4: Average Warranty Claims & Accrual Rates (as a % of product sales, 2003-2025)'. Accessed September 30, 2026.
LED Luminaire Lifetime: Recommendations for Testing and Reporting, Third Edition. U.S. Department of Energy Solid-State Lighting Program with the Next Generation Lighting Industry Alliance LED Systems Reliability Consortium (LSRC). published September 2014. Section 3 'Studies of failure modes', Figure 2 'The most commonly observed failures from LSRC member survey' (p. 11); key findings list (p. 14). Accessed September 30, 2026.
Little's Law as Viewed on Its 50th Anniversary. Operations Research, Vol. 59, No. 3, pp. 536-549 (INFORMS), J. D. C. Little, MIT Sloan; DOI 10.1287/opre.1110.0940. published 2011. Section 1 (Introduction), PDF p. 1: statement of L = λW; Section 4 (operations management), PDF p. 10: WIP = TH x CT (Hopp and Spearman); healthcare staffing discussion, PDF p. 11: queue growth as arrival rate approaches service rate. Accessed September 30, 2026.