CT reviews the listing — the price, the images, the seller history, the title and model number — and turns what it finds into a single 0–100 risk score and a BUY / VERIFY / BLOCK verdict. This page explains exactly how, what each signal means, and the limits of what a listing review can tell you. CT does not inspect, test, or authenticate the physical part.
Every check runs the same pipeline. Each signal is evaluated independently, marked PASS, WARN or FAIL, and given a weight. A FAIL on a high-weight signal moves the score far more than a WARN on a minor one.
The listed price is compared against the distribution of recent active and sold listings for the same manufacturer part number, in comparable condition.
How it scores: a price more than 1.3× the market median raises a flag; far above range (e.g. 4×+) is a FAIL. A price unusually below market is also flagged — too-good-to-be-true is its own risk pattern. The output names the median, the ratio, and the estimated overpay in currency.
The listing's images are matched against other indexed listings. If the same photo appears elsewhere — especially across listings with different model numbers — the photo may not depict the actual unit being sold.
How it scores: a single reuse is a WARN (request a photo of the real unit with serial/date code visible). The same image across many listings with conflicting models is a stronger signal. Reused imagery is not proof of fraud — it is a reason to verify before paying.
Account age, feedback volume and pattern, declared location, and whether the seller is an established industrial reseller or a new/low-history account.
How it scores: seller history adjusts confidence rather than driving the verdict alone. A strong price or image flag from a brand-new account weighs more than the same flag from a long-established reseller.
The manufacturer part number is checked against the listing title, the photographed label, and the stated condition for internal consistency.
How it scores: a title claiming one model while the MPN or label shows another is a FAIL — it is the most common way a buyer receives a non-matching component. Consistent MPN, title and label is a PASS.
Additional checks (condition-vs-photo consistency, cross-listing frequency, and listing-age patterns) feed the same score. The full breakdown for any given listing lists every signal with its result.
Each signal contributes a weighted amount to a composite score from 0 (highest risk) to 100 (lowest risk). The weights reflect how strongly each signal predicts a bad outcome — receiving a non-matching, non-working, or massively overpriced part. Price anomaly and MPN mismatch carry the most weight; seller history adjusts confidence.
The score is a structured summary of the listing signals — not a probability and not a guarantee. The same listing checked twice returns the same score; it is deterministic given the inputs available at check time.
The score maps to one of three verdicts. The verdict is advisory — it tells the buyer what posture to take before spending, not what the part physically is.
No material risk signals on the listing. Price in range, images and MPN consistent. Proceed with normal process.
One or more signals need confirmation before payment — request a unit photo, confirm the price, or check the seller.
Strong risk signals. Not ready for approval as listed. Resolve the flags or source from an approved supplier.
BLOCK does not mean the part is fake. It means the listing carries enough documented risk that approving it as-is would be hard to defend later.
Being explicit about the limits is part of the method. CT is a listing-review and documentation tool, not a testing lab.
What CT documents is the pre-purchase risk review and the approval decision around it — the signals seen, the verdict given, who approved, and when. This evidence may support an organization's ISO 9001 supplier-control (Clause 8.4) and NIS2 supply-chain governance processes. It does not, by itself, make a purchase compliant — that remains the buyer's and the organization's responsibility.
For monitoring customers, CT reports a Savings Score — an estimate of avoidable risk exposure, intended as a management metric, not an accounting figure. It is built from listings CT actually reviewed for that account.
The 20% factor is a deliberately conservative assumption of how much of the flagged spend would have become real loss (overpay, returns, downtime, replacement) had the listing been bought as-is. It is an estimate, labelled as such wherever it appears. CT does not claim the full flagged amount as "saved."
The Savings Score is only as meaningful as the volume reviewed — a few checks produce a noisy number. It becomes useful at portfolio scale, which is why it lives in the monitoring tier rather than per-record.
No listing-level screen is perfect, because the signal is the listing, not the part. Two failure modes matter, and naming them is part of using CT well.
A legitimate, fairly priced unit gets a VERIFY because its seller reused a stock manufacturer photo across several genuine listings. The fix is the same action CT recommends: ask for a photo of the actual unit. Verification clears it.
A counterfeit unit is listed at a normal price by an established account with consistent images and a correct MPN. The listing looks clean, so CT returns BUY — and the part is still bad. This is exactly why CT never claims to authenticate the part.
CT reduces the risk you can see in a listing. It cannot remove the risk you cannot see in a listing. The honest framing is "fewer bad buys and a documented decision," not "no bad buys."
Run a check on any industrial eBay listing. The verdict is free; the documented record is there when the purchase matters.
Check a listing →