01The starting point
Tourcrib was a brand-new touristic marketplace specifically catering to Martinique — a French island in the West Indies with a small but distinctive search market. No traffic. No content team. No brand awareness outside the island. The product was good; the discovery channel was empty.
We didn't have a content problem. We had a structure problem — and a positioning problem.— Founder note
02What we diagnosed
Three things surfaced once we mapped the search landscape:
Generic site structure that ignored hyperlocal intent
The marketplace was organized around product categories. Travelers search by location and activity type (‘rainforest hikes in Saint-Pierre,’ ‘Creole cooking class Fort-de-France’). The IA was fighting the way buyers actually look for things.
An unclaimed long-tail of district + activity queries
Hundreds of low-volume, high-intent queries — ‘best snorkeling spot Anses d'Arlet,’ ‘family activities Le François’ — were sitting unclaimed. Locally relevant, commercially valuable, almost no competition.
Zero authority signals
No backlinks, no editorial mentions, no schema. The brand was unknown to both Google and to the regional tourism ecosystem that drives those mentions.
03What we shipped
Hyperlocal information architecture
Built the site structure around Martinique-specific intent — locations, activity types, traveler personas — so each page targeted a real search instead of a generic category.
Programmatic local content
Shipped programmatic + hand-edited pages for every district, beach, hike, and restaurant cluster on the island. Each page answered one specific question a traveler types into Google.
Authority & link earning
Partnered with the regional tourism office, earned editorial mentions, and won a governmental innovation prize that translated into both signal and pipeline.
04The numbers, 18 months in
By month 18 the line was at 30K monthly organic visits and growing. The marketplace was the dominant channel for partner bookings, and the work earned a governmental innovation prize that compounded the brand signal.
05What didn't work
- Generic ‘top 10 things to do in Martinique’ listicles. They ranked, but converted poorly — travelers were already past that intent by the time they reached us.
- English-only versions of the landing pages. The audience was overwhelmingly French-speaking; the EN duplicates competed for the same intent without adding volume.
SEO at the hyperlocal level is fundamentally a structure-and-intent problem before it's a content problem. Get the IA right first, ship pages that match real queries, then earn the authority signals that make the ranking stick.