Connect commercetools with Elasticsearch to power fast, scalable catalog indexing, smart filters, and developer-tuned relevance, including synonyms and typo tolerance when needed.
• Product, variant, price, inventory, and category data is mapped from Commercetools to an Elasticsearch index schema, including localized attributes and channel-specific availability where applicable.
• Indexing is event-driven via Commercetools subscriptions or message queues, with delta syncs sending only changed entities and periodic full reindexing used to reconcile drift.
• Search documents are denormalized for read speed, and faceting fields are pre-shaped for filter and aggregation queries common in eCommerce search.
• Relevance configuration is applied in Elasticsearch through analyzers, field boosts, and query templates, with optional synonym sets and typo-tolerance via fuzziness depending on the chosen strategy.
• Ownership of truth remains in Commercetools, while Elasticsearch is treated as a derived read model; deletes, unpublishes, and attribute changes are propagated to keep the index consistent.
• Index operations, failures, and retries are logged and monitored, with dead-letter handling for malformed payloads or mapping conflicts to prevent silent data loss.
.png)
We stream product, category, price, and availability data from commercetools into a custom Elasticsearch index and expose it through a search API your storefront can call. Facets and filters are defined in the index mapping, so results stay consistent across channels.
Usually it’s SKUs, titles, descriptions, attributes, category paths, inventory signals, pricing, and search keywords, plus normalized fields for sorting and filtering. We also index computed fields like “in stock” and “best seller” when merchandising needs it.
Yes, but it needs deliberate analyzer setup, tokenization rules, and synonym management in Elasticsearch. Typo tolerance can be implemented with fuzziness and n-grams, then tuned so it helps discovery without polluting results.
We typically index localized fields per locale and market-specific fields for price, availability, and category rules. That lets you run per-market relevance tuning while keeping one integration pattern across regions.
It’s developer-led work, because mapping, analyzers, scoring, and reindexing strategy live in code and infrastructure. scandiweb has delivered 2,100+ eCommerce projects since 2003, so you get engineers who are used to owning critical performance layers like search.




