Methodology
Vendor scorecards are an honest transparency tool for shoppers — not a marketing tool for vendors, and not a lever you can pay for. Here is how they are built and how we keep them fair.
Own data only
Every metric is aggregated from this vendor’s own crawled data alone (catalog breadth, data completeness, freshness, price positioning). The only cross-vendor numbers are the public market median and range every shopper sees. We never invent a quality, trust, or reputation score we lack the data for.
Coverage-confound suppression
We distinguish “this vendor lists little” from “we’ve crawled little of this vendor.” Every scorecard states our coverage and freshness; if our coverage is thin or stale, the scorecard says so and suppresses comparative verdicts rather than penalising the vendor for our gap. Thin data means “not enough data,” never a low grade by default. A small vendor is never punished for being small.
Never pay-to-rank or pay-to-improve
No metric has any paid input; the computation reads only real crawled fields. There is no single overall grade to sell or game, and no “claim your page to fix your score” upsell. Claiming (a merchant account) unlocks private detail and a correction channel — never a better public score. The scorecard is informational and does not affect search ranking.
Right of reply & corrections
If a vendor sees a factual error (wrong price, miscrawl, dedup error), they can flag it; it routes to our data-quality review and is corrected honestly — never bought. A claimed merchant may also post a public response shown on their page.
Supply is not demand
Scorecards describe catalog and price facts (supply). We do not publish per-vendor demand or conversion, and we do not rank vendors by popularity.
This methodology is part of our anti-manipulation policy: ranking and trust are never for sale.