companies of all sizes
Cross-source entity resolution: matching the same product everywhere
How Scrapliy knows a product on your store, a competitor's site, and a marketplace feed is the same item.
Updated 2026-04-05 · Platform · 8 min read
Step-by-step walkthrough
Identity keys, strongest first
A validated GTIN/EAN/UPC (checksum-verified, not just digit-count) is the strongest match. Without one, brand+MPN or brand+SKU is used. As a last resort, a fingerprint combining brand, normalized name, and image hash provides a soft match.
Soft matching when there's no strong identifier
Name similarity combines word-set (Jaccard) and word-order-insensitive character similarity, with a penalty when numeric tokens differ—so "Air Max 90" and "Air Max 95" don't get merged just because they're one character apart.
What you get once products are resolved
One product identity aggregates every source's price and stock, with automatic conflict detection (including cross-currency, normalized via daily exchange rates) and full field history over time—not just the latest snapshot.
Ready to set this up for your store?
Add a site, pick a crawl schedule, and Scrapliy handles discovery, extraction, and change detection from there.
Related: Platform → · Compare tools · Blog · All guides