Problem
A parts store must show only products that fit the shopper's vehicle or equipment, and fit depends on several attributes at once. If a SKU keeps Year, Make and Model as separate lists, nothing records which values belong together. A filter can then accept combinations that do not exist, especially when one SKU fits many vehicles. Merchants also use different dimensions, so a fixed Year/Make/Model schema does not carry over.
Example from the article: a product lists 2018, Ford, Honda, F-150 and Civic. Every value is valid for the SKU, so a naive filter treats a "2018 Honda F-150" as a match even though that vehicle does not exist.
Solution in brief
Store each valid combination of fitment attributes as its own fitment entity, indexed like a product or collection and linked to the products it applies to. Where no entity exists, the combination is impossible. The filter, search, product page, cart and chat assistant all query the same entities, so none of them implements its own fitment logic.
Questions this case answers
- How to model multi-attribute compatibility between catalog items and target objects so that filters never produce combinations absent from the source data
- How to keep one definition of compatibility shared by several interfaces (filters, query box, item page, cart, conversational agent) in item search
- How to keep a vehicle-compatibility (fitment) filter from returning impossible Year/Make/Model combinations when one SKU fits many vehicles?
- How to support different fitment schemas (Year/Make/Model, Make/Year/Engine, marine) in one parts catalog system?
Tools and methods
- Fitment entity model — one entity per valid attribute combination, dimensions set per merchant
- Algolia index — holds products and fitment entities for filtering and search
- Synonyms — map shopper terms (engine family, abbreviation, OEM number) to catalog terms
- Shopify — operational system; entities attach to its products, collections, customers and cart lines
Google or a one-time email code. No password.