stalance/services/ecommerce-engineering
Catalogues that stay clean at millions of listings.
Most catalogue problems are identity problems: the same product three times, twelve shades as twelve listings, a category tree that grew by accident.
- in production
- For our own 20+ storefronts
- scale
- 4,000,000+ product listings
- merchants
- 50+ merchant programmes
- history
- A price log on every variant
What we build
Identity and de-duplicationThe same product from several sources becomes one record, resolved at ingest rather than patched later.
Variants and attributesColour, size and material as structured values, with a multilingual vocabulary for colour.
Faceted searchFilters that match the catalogue’s real shape, not a generic template, with counts that add up.
Category trees per verticalA taxonomy per market, mapped from messy source trees, with empty branches removed.
Price historyA change log per variant, so “this price just dropped” is a query, not a guess.
Structured data and technical SEOSitemaps, canonicals, structured data and redirects for products that moved or vanished.
How we would approach yours
01 auditAudit the catalogueIdentity collisions, empty branches, attribute fill rates. A short written report.02 modelFix the modelVariants, attributes and categories as data, with the rules written down.03 searchRebuild search on itFacets and ranking from the cleaned model, with counts you can verify.04 seoProtect the URLsRedirects, canonicals and sitemaps so the move does not cost rankings.Questions engineers ask us
Which platform do you work with?
Usually your existing one. We pick tools per problem and build on what is a reasonable choice.
Can you clean a catalogue without replatforming?
Yes. Identity and attribute work is mostly a data problem and can ship behind the current storefront.
Do you handle marketplace and merchant-feed integrations?
Yes, and we run them ourselves, so we know where the feeds lie.
Tell us what is broken
One paragraph is enough. An engineer reads it, not a ticket bot.