Customer storyGédimat
Local products published nationwide, and product data on a new scale
At Gédimat, each member enriches its own SKUs to publish them on the national website. Seeing this model work, the network gave Upsellr a second project: enriching and restructuring its national product database.
At a glance
Gédimat is a national network of independent merchants in building materials, DIY and outdoor landscaping. Six members use Upsellr to enrich around 1,000 local products each and publish them to the national website’s standard. In parallel, Upsellr enriched nearly 90,000 products in the national database, and took the product repository from 400 to around 1,500 attributes used as filters.
- Each member keeps control of the local products it publishes.
- Around 1,500 attributes kept as filters, out of more than 80,000 identified.
- A new category tree, for more precise search.
- gedimat.fr, the first source cited by ChatGPT on the local products tracked.
A network of independents, two projects
Gédimat is a national network of independent merchants in building materials, DIY and outdoor landscaping. Each member runs its own store, and the product project is led at network level.
The first project covers local products: members list products of their own and want to show them on the national website, to the standard it requires. Seeing this model work led the network to launch the second: enriching its national product database.
The goal
Publish local products nationwide, then give the network’s product data a new lease of life.
Without taking control of their own products away from members.
Each member publishes its products to the national standard
Each member has its own Upsellr account and regularly submits the EANs of its products. The agents enrich them, generate the content and bring them up to the national standard before publication on the Gédimat e-commerce website.
- active members, more being rolled out
- 6
- local products enriched per member
- ≈ 1,000
On these products already online, Gédimat is highly visible in ChatGPT’s answers: gedimat.fr is the first source used, ahead of Leroy Merlin and Point P.
- of tracked products cited at least once
- 91%
- average presence in answers
- 63%
- first-position citations, versus 214 for Leroy Merlin
- 418
First Upsellr measurement on September 1, 2026, on ChatGPT, 495 products and about 3,000 queries. It is a starting point, not yet a trend.
A national database on a new scale
On the national database, Upsellr first automatically classified the products in the existing taxonomy, then supported the construction of a new category tree, for more precise product search.
At the start400
With Upsellr≈ 1,500
More than 80,000 potential attributes were identified, then sorted: about 1,500 were kept and filled in. Nearly 90,000 enriched products are being integrated into the PIM, with the Gédimat teams.
Conversion and visibility cannot be measured on this scope yet: the products are not published yet.
90,000 products enriched and exported in our national PIM format: Upsellr let us handle a volume impossible to process by hand, with human review at every step.
Nicolas GauguezHead of Digital Performance, GédimatWhat comes next
- measure how long it takes to publish a local product, before and after Upsellr;
- complete the integration of the 90,000 products into the PIM, then measure conversion and visibility;
- extend GEO measurement to Gemini, Perplexity and Google AI Overviews.
Sources and method
- Volumes and attributes: Upsellr platform data; volume exported to the national PIM confirmed by Gédimat.
- GEO visibility: first Upsellr measurement on September 1, 2026, on one engine (ChatGPT), 495 products and about 3,000 queries.
Marcus
Amazon