Thousands of spellings per product. One purchase intent. Zero manual rules.
How the camping specialist automated its long-tail on the product discovery technology it already ran — and lifted search conversion by 67%. No engine swap. No extra headcount.
Camping depth that became an online liability.
Fritz Berger is one of the largest camping specialists in the German-speaking market: over 100 stores, nearly 800 employees, and a range spanning tent pegs to camping stoves to outdoor apparel — family-owned for 68 years and now expanding internationally.
Online, every category brings its own jargon, dialects and spelling chaos: customers search for the same item with hundreds of different terms. The product depth that makes Fritz Berger strong in brick-and-mortar became an Achilles’ heel in ecommerce.
Manual optimization hit a wall.
Fritz Berger has run a solid search technology since 2022 — smart ranking, flexible merchandising, well-structured data enrichment. The top 1,000 search terms performed well. The problem lived deeper, in the long-tail: hundreds of spelling variants per product group, terms searched once and never again, and zero-result pages that frustrated customers — and even kept store staff from finding products in the system.
E-commerce specialist Sonja Genitheim spent her days on synonym maintenance instead of strategic category work. New products couldn’t be search-optimized fast enough. And the planned international expansion threatened to multiply the same problem in every new language.
An optimization layer in front of the existing search.
searchHub doesn’t replace search technology. It sits in front of it — between the customer and the engine — and fixes what reaches the search, not the search itself.
“Tent carpet”, “outdoor rug”, “awning mat” and hundreds of variants more: one purchase intent. searchHub groups them into a single cluster, based on which terms actually led to clicks and purchases.
From each cluster, searchHub elects the statistically strongest term as the ambassador or master query. Only that optimized query reaches the engine. Fewer queries, better results.
New spellings are detected and assigned to existing clusters without manual intervention. The system learns from real customer behavior, continuously.
Autocomplete becomes intent prediction: suggestions ranked by actual purchase probability, not alphabetical order.
The existing technology keeps its strengths — ranking, merchandising, data enrichment. searchHub adds what was missing: automated language understanding for the long-tail. Or, as Head of E-Commerce Christoph Kirner puts it: “This isn’t a weakness of the existing search technology. The search terms for a single product are simply too varied. No human can manually maintain hundreds, or in some cases thousands, of spelling variants.”
Numbers that don’t need a sales pitch.
Integration took a few weeks. The A/B test started in August 2024; full rollout followed on October 15th. The results beat every expectation.
“Even if I manually optimized 20% of the terms, I’d never match what searchHub delivers on its own,” says Sonja Genitheim. The outlook holds, too: as the international expansion continues, country teams run the search technology for marketing optimizations while searchHub runs on its own in the background. The integration is built to be reused — what works once, works everywhere. searchHub isn’t a project. It’s infrastructure.
Groups long-tail keywords into purchase-intent clusters, so only the strongest query per cluster reaches the engine — and autocomplete turns into intent prediction: proven, high-converting suggestions ranked by actual performance, not alphabetical order.
How we solve it