ecommerce_ops · ecommerce · workflow
Delivery Hero builds semantic product matching using retrieval-rerank and hard negative sampling
Delivery Hero needed an algorithm to match product titles across its own catalog and competitors' catalogs for pricing strategy and assortment gap analysis, and to detect internal duplicate items.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · Product title query submitted
A product title p is submitted to find matching titles within an unordered set S of product or competitor titles.
Tools used
SBERTLucene
Outcome
The Retrieval-Rerank approach augmented by hard negative sampling enables effective identification of similar products and assortment gap management while balancing computational efficiency and accuracy.
What failed first
Lexical matching could not recognize the same product described with different words — unit abbreviations, misspellings, and missing words all produced false non-matches.
Results
Volumesignificantly higher accuracy
Grounding & classification
Source type: technical build writeup
12 fields verified against source quotes.
enterprise searchproduct catalogfailure mode describedtools describedworkflow describedecommerceaccuracy improvementtechnical build writeupecommerce ops