Supplier Scorecard for Cable Assembly Buyers

Agree the definitions before the metrics

A metric is a name plus a definition, and the name is the part everyone agrees on. Settle these first, in writing, or the scorecard becomes a dispute about arithmetic.

Question	Why it changes the number
On time against which date?	The date you requested, or the date the supplier confirmed. Suppliers usually score against the confirmed date, buyers care about the requested one, and the gap between them is where expediting lives
On time at which event?	Ship date, or arrival. A supplier controls one of those
What is the window?	Early is often as disruptive as late for a line feeding a build sequence. State whether early counts as on time
What counts as a defect?	Per unit, per lot, or per line item. A rejected lot of 500 is one event and 500 units, and those are different rates
Whose fault is a hold?	A shipment late because a drawing question went unanswered for a week is a shared failure. Decide in advance whether it scores

Four areas worth scoring

Keep it short. A scorecard nobody completes is worth nothing, and a long one is a scorecard nobody completes.

Delivery. On-time performance against the agreed definition, plus the size of the miss when it misses. A supplier that is late by a day occasionally is a different risk from one that is late by three weeks rarely, and a single percentage hides that.
Quality. Reject rate on the agreed basis, plus escapes — nonconformances found after acceptance rather than at receiving. Escapes are the number that predicts trouble, because they say the supplier's own controls did not catch it.
Responsiveness. Time to acknowledge an RFQ, time to a firm date, time to answer a technical question. This is the area that most reliably predicts how a program will feel to run, and it is almost never scored.
Documentation. Whether certificates, test data and traceability records arrive complete with the shipment, or after three requests. See cable assembly testing for what the test records should contain and counterfeit parts avoidance for why the traceability half matters.

Corrective action is the real signal

Any supplier will have a bad month. What separates them is what happens next, and it is worth a line of its own on the scorecard: how long a corrective action takes to close, and whether the same finding comes back.

A repeat of a closed finding is the single most informative event in a supplier relationship. It says the corrective action addressed the symptom, and it is a better predictor than any delivery percentage.

Make it a conversation, not a verdict

Share the raw data, not just the score. A supplier cannot fix a number it cannot reconstruct.
Score a period long enough to mean something. Monthly scoring on four shipments a month measures noise.
Say what the score is used for. If it feeds sourcing decisions, say so; if it does not, do not imply otherwise.
Let the supplier dispute a data point and record the outcome. Most disputes are definitional, and each one resolved improves the next quarter's data.

What it cannot tell you

A scorecard is a record of a relationship that already exists. It says nothing about a supplier you have not used, which is what the supplier audit checklist and verifying certifications are for, and it does not capture capability — a supplier can score well on simple assemblies and be the wrong choice for a harness with circumferential shield terminations and a controlled marking process.

It also cannot substitute for a clear package. Much of what looks like supplier underperformance traces back to an incomplete drawing or an RFQ that left the supplier to assume — see harness drawing requirements and the RFQ checklist. Scoring a supplier on ambiguity you introduced is a way of measuring your own process and calling it theirs.