Elasticsearch, tuned for B2B. The architecture underneath, untouched.
SLV relies on open-source and expert software for onsite search: the Open Commerce Search Stack gives Elasticsearch robust variant handling inside Spryker, and searchHub smartQuery sends it the best-performing query. 11 languages, 15 countries, live within weeks.
Professional lighting — from Übach-Palenberg to more than 100 countries.
SLV is a leading supplier of light fixtures and lighting solutions, and a B2B brand through and through: its products light private homes, hotels and restaurants, offices and commercial premises — specified by electricians, lighting designers and architects. The promise is a balance of innovative technology, modern design and a price-performance ratio the trade counts on, with service to match.
Inspired by German engineering, SLV grew from its roots in Übach-Palenberg near Aachen into 15 country subsidiaries; from an 18,500 m² central warehouse, fittings and lamps ship worldwide every day. The BIG WHITE®, its annual catalog, is an industry standard — and online, all of that depth has to be searchable in 11 languages.
Make Elasticsearch perform for B2B — inside Spryker, without leaving the standard.
Elasticsearch is close to the industry standard for search, and it ships integrated with the Spryker shop system SLV runs. But its ecommerce search has known gaps, and a professional catalog exposes every one of them: one product in many colors, materials and sizes that Elasticsearch, out of the box, doesn’t know how to group; buyers who search in 11 languages, with the spellings and typos that come with them. The built-in search didn’t meet all client expectations.
The requirement was firm: stay close to the standard Elasticsearch integration in Spryker — so the platform stays updatable and the business keeps running — while meeting the demands of 15 countries and 11 languages.
A tuning kit for Elasticsearch — and an agent in front of it.
Nothing underneath changed. Elasticsearch still runs in the background of Spryker; the complexity moved into two layers on top of it — one open source, one expert software. Top to bottom, this is the path a query takes.
In front of everything. smartQuery clusters the variations of a search term — spellings, typos, word order, in all 11 languages — by purchase intent, and sends only the best-performing query of each cluster on to the engine. It runs continuously on real search data; every decision stays SLV’s to override.
From the team behind searchHub. OCSS takes over what Elasticsearch handles poorly out of the box for ecommerce: robust handling of product variants — color, material, size — so a product shows up once, in the right version; flexible facets and search strategies with strong presets; fallback queries that relax a query when the first attempt returns nothing.
Still the engine, still the standard Spryker integration. Updatable, flexible, business as usual — searchHub works the same way in front of any Elasticsearch.
How searchHub works with ElasticsearchImplementation took a few weeks, in close cooperation between SLV, searchHub and the agency in charge, best it — which calls the combination “a veritable recipe for success” and plans to use OCSS + searchHub in several other search optimization projects.
Thousands of queries a day. One search experience in 11 languages.
searchHub now optimizes thousands of SLV customer queries every day: long-tail searches are grouped into clusters of equal intent, and from each cluster the result with the highest sales is selected. Variant search works the way a professional expects — by free-text entry or by facets, buyers land on the right color, material and size quickly. And because the optimization runs continuously, results keep improving on the fly, with no change to Elasticsearch or Spryker.
11 languages served from the same Elasticsearch and the same Spryker standard.
Color, material and size resolve to one product — by free text or by facets.
New spellings and typos are clustered as they appear. The team overrides; it doesn’t maintain lists.
Groups long-tail keywords into purchase-intent clusters — spellings, typos and word order, in every language — so only the strongest query per cluster reaches Elasticsearch. Search managers keep full control.
How we solve itOpen-source Elasticsearch for ecommerce, maintained by the team behind searchHub: optimized indexing, variant handling, relevance, filtering and faceting, search management — a solid default configuration, API-first, in front of the Elasticsearch you already run.