Buyer questionAnswered from the China side
Supplier scorecard: the one row you cannot fill from your own data.
A scorecard works by turning things you already know into numbers you can compare. Quality, delivery, cost and responsiveness all come from your own transaction history, which is why they are the easy rows — and why a scorecard is close to useless for a supplier you have not bought from yet. There is one dimension that never comes from your history, is usually labelled “compliance” or “risk”, and is the row most often left blank, copied from the supplier’s own claims, or quietly scored as 5/5 because nothing bad has happened yet.
SUPPLIER SCORECARD / 01
The scorecard is only as honest as its weakest row.
This page is about one row. It does not supply a template, and it does not score performance — only you have that data.
The four dimensions you already own
Most scorecards, whatever the layout, measure the same four things from data that sits in your own systems:
Quality — defect rate, rejections at incoming inspection, warranty returns. Delivery — on-time rate against confirmed dates, lead-time variance. Cost — price movement, quotation accuracy, hidden charges appearing late. Service — response time, how problems are handled once they exist.
These are backward-looking by construction. They are exactly what you want for a supplier you have run twenty orders through, and exactly what you do not have for the supplier you are evaluating this week. A scorecard built only from these four cannot be filled in at all before the first order — which is the moment when the decision actually carries the most risk.
The fifth row, and why it gets guessed
The fifth dimension goes by several names — compliance, risk, governance, supplier legitimacy. Unlike the other four it is not about performance; it asks whether the counterparty is a real, currently active legal entity, registered for the work you are buying, with nothing adverse on its public record.
It gets guessed for a structural reason: there is no internal data to score it from. So the row is filled from whatever the supplier provided — a licence photograph, a certificate PDF, a platform badge. Each of those is a document the supplier chose to send. Scoring a supplier on documents it selected is not verification; it is transcription.
The failure mode is not that the row is scored badly. It is that it is scored confidently, from evidence that was never independent.
What belongs in that row for a Chinese supplier
Each item below resolves to a fact with a source and a query date, not to an impression:
Registered name and Unified Social Credit Code resolve to one entity. The 18-character code is the identifier; an English trading name is not registered and cannot be checked. How the check is run.
Registration status is active. Struck off, revoked, or in liquidation are all states a website continues to look perfectly normal in.
The registered scope covers what you are buying — and in particular whether it covers manufacturing at all. Reading manufacturer versus trader from the scope.
Business-abnormality listings, administrative penalties, former names. Counts first, detail second. A former name is not a red flag by itself; a former name that appears between your quotation and your purchase order is a question.
The payment beneficiary matches the contracting entity. Strictly this belongs to the payment step rather than the scorecard, but it is the check with the highest value per minute spent. How to compare the beneficiary against the register.
Scoring it without inventing precision
Do not convert “could not be read” into a number. When a source is unreachable or a field is not returned, the honest entry is unknown. Scoring it zero punishes a supplier for our access problem; scoring it full marks manufactures a clean result out of nothing.
Distinguish “no record found” from “no record exists”. Commercial platforms republish official disclosures with a lag. A platform showing nothing on a query date is evidence about that platform on that date, and nothing stronger.
Date every entry. A registration status is a snapshot. Six months later it is history, not a fact, and re-checking before a large order costs minutes.
Weight it as a gate, not a score. The identity items are pass/fail in practice — a supplier whose name and code do not resolve to one entity does not get a 2/5 on legitimacy and a good weighted average. It gets a stop.
What this desk does not provide
Stated plainly so the page is not read as an offer it cannot honour:
No scorecard template or software. Plenty of those exist and they are not our work. What we produce is the evidence for one row of yours.
No performance rating. We have no visibility of your defect rates, your delivery dates or your prices, and any number we attached to them would be fabricated.
No single supplier score. A composite number would hide exactly the thing that matters — which item failed, and against which source on which date.
We verify records and report them with their source, retrieval date and limits. We do not certify that a supplier is safe, approved, or good to work with — those are conclusions the record cannot support.
Filling the row for a supplier you are looking at now
You need their exact registered Chinese name, ideally taken off a photograph of the business licence (营业执照). With that, the identity and standing items above can be checked directly.
The free check returns registration facts and public-signal counts. Detail behind those counts, and a written report you can attach to a supplier file, are the paid tiers.