Product matching

Know when two listings are the same product.

Crawlr resolves inconsistent titles, identifiers, variants, units, and pack sizes into dependable product relationships. Every match keeps enough evidence for analysts to understand, review, and trust the result.

See the matching workflow

Evidence, not guesses

Every product relationship keeps its reasoning.

Exact identifiers, normalized attributes, pack logic, variant evidence, and conflicting signals remain attached to each proposed match.

Canonical identifier
EXACT
Pack and unit
ALIGNED
Confidence evidence
REVIEWABLE

How matching works

From noisy catalogs to a dependable identity layer.

  1. 01

    Normalize the evidence

    Standardize identifiers, brands, attributes, variants, units, and pack sizes across source listings.

  2. 02

    Resolve and score

    Combine exact and normalized evidence while separating conflicts from confident relationships.

  3. 03

    Deliver with review state

    Return matched groups, confidence evidence, and a clear queue for ambiguous candidates.

Operating detail

What goes into the workflow—and what comes out.

Capabilities

  • Exact identifier and canonical attribute matching
  • Variant, pack-size, and unit normalization
  • Evidence and confidence attached to every relationship
  • Review queues for ambiguous or conflicting candidates

Inputs

  • Source listings
  • Reference catalog
  • Identity rules
  • Review thresholds

Outputs

  • Matched product groups
  • Normalized attributes
  • Confidence evidence
  • Needs-review queue

Use cases

  • Price comparison
  • Catalog onboarding
  • Marketplace analysis
  • Competitive assortment

Delivery modes

  • API
  • Matched files
  • Integrated project delivery
  • Private client web app

Product matching FAQ

Answers about matching, evidence, and review.