A supplier scorecard is meant to make a supplier comparable quarter to quarter. Most of them end up measuring what is easy to collect: unit price variance, on-time delivery, and a quality figure that comes from the supplier's own report. A supplier can score in the high nineties on that set and still ship a relabelled part, change a die without telling anyone, or deliver a kit that is one reel short. The score looks fine because the questions were never asked.
The fix is not a longer form. It is a smaller set of measures, each with a data source the buyer can check, weighted so that the failures that actually stop a line can move the number.
Where scorecards lose their value
The first failure is measuring at purchase order level. A PO can be on time while the kit inside it was complete only because two lines were back ordered and shipped a week later. Measure the line item, not the order.
The second is accepting supplier-reported quality without a route to verify it. If the first pass yield comes from the same system that the supplier uses to decide what to report, it is a claim, not a measurement.
The third is ignoring provenance. On a component route, the question is not whether the part worked, it is whether the part is what the label says. X-ray counting, marking comparison and date code checks are the only evidence that answers it, and a scorecard without a line for them rewards the wrong behaviour.
The fourth is a score with no trigger. A number that nobody acts on is a report. The scorecard needs a floor and a consequence attached before the first review, otherwise the first bad quarter is spent arguing about the weighting.
The fifth is weighting everything equally. Late paperwork and a wrong part are not the same event, and a scorecard that treats them the same way tells a buyer nothing about which supplier to keep.
A metric set that holds up
| Metric | How to measure it | Source the buyer can audit | Weight guide |
|---|---|---|---|
| Line complete on time | Kit lines delivered complete, against the promised date | Kit document against the schedule | High |
| Kit accuracy | Lines short or wrong per hundred issued | Shortage report by work order | High |
| First pass yield | Boards passing ICT and functional test first time | Test log by lot | High |
| Provenance evidence | Lots with X-ray count, marking and date code record | Inspection record per lot | High |
| Traceability response | Hours to answer a lot to board question | Sample request, timed | Medium |
| Change notice lead time | Days between notice and first affected shipment | Change log and shipping dates | Medium |
| Corrective action closure | Closed with an effectiveness check, not a ticket | Action log plus a re-audit result | Medium |
| Date code at receipt | Age of material on arrival, by part class | Receiving record | Medium |
| Quote response | Working days to a quote, including a quantity change | Quote log | Low |
| End-of-life notice | Notice before the last order date, in writing | Change log against the maker notice | Low |
Ten lines are enough for a quarterly review. Beyond that, the data collection costs more than the insight, and the supplier learns to manage the form instead of the process.
An assembly route and a component route score differently
The same template applied to both hides what matters. On an assembly route, first pass yield, kit accuracy, process data per lot and corrective action carry most of the weight, because the output is boards and the evidence is in SPI, AOI and X-ray records. On a component route, provenance evidence, date code age and change notice carry the weight, because the output is a part whose history decides whether the board is what the customer approved. Where one supplier does both, score the two scopes separately and do not average them into one figure.
Two measures are worth adding for programs with a long service life: how fast the supplier answers an end-of-life question, and whether the answer arrives as a written quantity basis. Those two predict the cost of the last-time buy years before it is placed, which is set out on last-time-buy and LTB planning.
What SUPERB PCBA can put in front of you
The evidence is per lot rather than per quarter. Components are counted by X-ray on receipt, markings and date codes are compared with maker data, parts that carry provenance risk are opened and checked, and aged stock is verified for solderability before release. Moisture sensitive parts are baked at 125 C to J-STD-033 and stored under ESD S20.20 in an SMT area at ISO 8, 22 plus or minus 3 C and 45 to 65 percent relative humidity.
On the build side, seven SMT lines place down to 01005, with 3D SPI on paste volume, 3D AOI plus DIP AOI, X-ray in 2D and CT to IPC-7095 with voiding below 5 percent, twelve zone reflow on SAC305 with nitrogen below 500 ppm, ICT and functional test, across seventeen process steps and ten QC stations under IPC-A-610 Class 2 and Class 3, ISO 9001, and ISO 13485 for medical programs. Verified lots carry a 730 day warranty, and every pin to pin change is made only with written approval.
The systems behind those numbers are the ones a buyer can read before the first order: the audit scope is described on supplier quality audit, the sampling levels on incoming inspection plan, and and the pre-shipment check on PCBA inspection before shipment.
Supplier scorecard questions
How often should the scorecard be reviewed?
Quarterly with a monthly data feed. Quarterly scoring with no monthly data leaves a full quarter of bad lots to be discovered at the review, when the build that used them has already shipped.
Should the supplier see the raw data?
Yes, and the same data the buyer holds. A scorecard that is discussed but never shown turns the review into a negotiation about perception, and the supplier cannot correct a trend it cannot see.
What is a sensible floor for a new supplier?
Set it before the first order and attach it to volume, not to opinion: a score below the floor for two quarters triggers a review and a hold on new work. The number matters less than the consequence being written down in advance.
Can a scorecard replace an audit?
No. The scorecard shows the trend and the audit shows the process. The audit is where you see whether a lot can be followed from receiving to the line, which is the one question a report cannot answer.
Next step
Send the scorecard template your team already uses. Each line comes back with the data behind it and a reference to the record it came from, and any line that cannot be evidenced is marked as such rather than filled with an estimate, which is the only version of the number worth reviewing.