Implantable Electronics KPIs: Seal Yield, Scrap, Labor, Supplier and Recall Benchmarks

Use a benchmark as a first reference point, not as a target. In implantable devices one scrapped unit can erase the margin on ten good ones, so the ranges below are most useful for finding the gap that costs you most, not for grading the plant.

Every figure here shows its basis. A superscript number links to a public source in the Sources list at the end, and the sentence states that source's context. Most figures are marked no cited source: they are kept exactly as originally published because none of the public documents reviewed for this guide states them, so treat them as illustrative.

At a glance

Read the scoreboard as a set

Implantable device plants live and die on a short scoreboard: rolled throughput yield, hermetic seal pass rate, scrap rate by value, cost of poor quality, cleanroom labor efficiency, test station utilization, documentation right-first-time, and mock recall speed. Typical and world-class performers separate by a factor of 2 to 5 on most of these no cited source.

The ranges in this guide are meant for active implantable manufacturers running 2,000 to 50,000 units per year no cited source. Measure each KPI the same way every month before chasing any of them.

Rolled throughput yield and hermetic seal yield

Rolled throughput yield across the full build, assembly through final test, runs 80 to 88 percent at typical plants and 94 to 97 percent at world-class ones no cited source.

The dominant sub-metric is hermetic seal first-pass rate. Typical is 95 to 97 percent, and world class is 99 percent or better on laser-welded titanium cans no cited source.

Seal failures concentrate in a few causes: 60 to 70 percent usually trace to weld parameter drift or feedthrough braze voids no cited source. Both are fixable with tighter incoming checks, worth 1 to 3 yield points in 90 days no cited source.

How to measure

• Track seal yield weekly on a rolling 4 week window so 100 unit lots do not whipsaw the trend no cited source.
• Define the yield denominator as units started, not units tested, or seal yield inflates 1 to 2 points when gross leak rejects get pulled before fine leak no cited source.
• Use rolling 4 to 8 week windows for anything with lot sizes under 200 no cited source.

What moves it

• A Pareto of seal failures, worked from the top cause down.
• Weld parameter control that stops drift.
• Tighter incoming checks on feedthrough brazes.
• A scoreboard posted at the weld station, which typically gains 1 to 2 points within a quarter from operator attention alone no cited source.

Track seal yield with the Hermetic Seal Yield calculator, and size weld inspection with the Micro-Weld Inspection Load calculator.

Final electrical test: first-fail rate and station utilization

First-fail rate at final electrical test is the second yield watch item. A rate of 4 to 8 percent is typical, and under 2 percent is world class no cited source.

Test station utilization runs 55 to 70 percent typical against 80 to 85 percent world class no cited source. Before adding capacity, find whether the constraint is program length, retest volume, or scheduling gaps, because adding a second shift is usually the wrong first move.

For example, if 5 percent of units fail electrical test, moving that screen one station earlier, before encapsulation, saves the full battery and case value on every one of those catches.

How to measure

• Log retests separately from first passes.
• Count a device that passes on a retest as a fail for KPI purposes, since passing on attempt two consumed double the station time.
• Split station time into program length, retest time and idle time.

What moves it

• Electrical screening moved ahead of encapsulation.
• Shorter test programs.
• Lower retest volume.
• Closing scheduling gaps before adding shifts.

Model test and other station loads with the implantable electronics calculators.

Scrap rate by value and cost of poor quality

Scrap rate by value, not by unit count, is the honest measure, because a reject at encapsulation costs 10 times one at kitting no cited source. Typical plants scrap 3 to 6 percent of material value, and world class holds under 1.5 percent no cited source.

Cost of poor quality (COPQ) means scrap plus rework plus complaint handling plus audit findings. It runs 10 to 15 percent of revenue at typical medical device operations and under 5 percent at the best no cited source. If more than half your COPQ sits in the last two operations, your inspection points are too late in the flow no cited source.

How to measure

• Weight scrap by accumulated standard cost at the failing operation.
• Refresh those standard costs quarterly.
• Feed monthly reject codes by build stage into the scrap calculation.
• Add rework, complaint handling and audit findings to scrap for COPQ.

What moves it

• Inspection points moved earlier in the flow.
• Electrical screening before encapsulation, where rejects cost most.
• Fewer seal failures, the dominant yield loss.

Cleanroom labor efficiency

Cleanroom labor efficiency is earned standard hours over paid hours. It runs 72 to 82 percent typical and 88 to 92 percent world class no cited source, and the gap is mostly gowning discipline, kit completeness, and batch sizing.

Compare predicted hours against actual payroll hours each month. A gap above 10 percent means your standards are stale no cited source.

How to measure

• Divide earned standard hours by paid hours.
• Compare predicted hours with actual payroll hours every month.
• Refresh the labor standards when the two drift apart.

What moves it

• Gowning flow and material staging, worth 5 plus efficiency points without headcount no cited source.
• Kit completeness at the start of each build.
• Batch sizing.

Build predicted hours with the Cleanroom Assembly Labor calculator.

Supplier incoming defect rate

Supplier and documentation KPIs decide audit outcomes. Incoming defect rates from implant-grade suppliers run 500 to 2,000 PPM typical no cited source, and world class is under 100 PPM¹, with skip-lot status earned by history.

The cited threshold comes from aerospace and defense, not implants. In its 2026 supplier rating criteria, RTX puts high-volume suppliers under 100 PPM in its top Performing tier, up to 250 PPM in Progressing, and above 250 PPM in Underperforming¹. At that manufacturer, under 100 PPM is the entry point to the top tier rather than an exceptional result¹.

How to measure

• Count defects per million parts received, by supplier.
• Include failures found on the floor as well as at incoming inspection, as RTX does¹.
• Grant skip-lot status only on a documented history.

What moves it

• Requalifying critical suppliers every 12 to 24 months, not every 36, as best performers do no cited source.
• Tighter incoming checks on the parts behind seal and weld failures.

Device history records and mock recall speed

Device history record (DHR) right-first-time is 85 to 93 percent typical and 98 percent or better world class no cited source. Every rejected DHR adds 20 to 45 minutes of review labor no cited source.

Mock recall traceability, from order to full genealogy report, is 4 to 8 hours typical and under 1 hour with an electronic DHR (eDHR) no cited source.

Recall speed has a regulatory clock in the U.S. A device maker that starts a correction or removal to reduce a risk to health must report it to FDA within 10 working days of starting it², so genealogy has to be reportable well inside that window.

How to measure

• Count DHRs accepted at first review over DHRs submitted.
• Convert rejected DHRs into review hours and full-time equivalents.
• Time mock recalls from the order to the full genealogy report.

What moves it

• Converting paper DHRs to electronic, which cuts documentation load 50 to 70 percent no cited source.
• Complete genealogy captured at each operation, not reconstructed afterward.

Work the levers in payback order

Levers, fastest payback first

• Pareto the seal failures and tighten incoming checks on welds and feedthrough brazes.
• Move electrical screening before encapsulation.
• Attack gowning flow and material staging.
• Convert paper DHRs to electronic.

Rerun the matching calculation after each change. A lever that has not moved its KPI within 8 weeks is not working no cited source.

How to build your own baseline

• Write down each KPI's definition and denominator and keep them fixed.
• Measure every KPI the same way each month before chasing any of them.
• Weight scrap and yield losses by the stage where they occur.
• Compare against your own trend first and published ranges second.
• Post the scoreboard weekly on the floor and note what changed whenever a KPI moves.