The story
Dario runs an auto parts store with tens of thousands of SKUs. The defining property of every product is what it fits, and the defining property of every search is a vehicle — a year, a make, a model, an engine variant, or an OEM number that encodes all four.
His fitment data existed in a lookup widget that ran on the page and was invisible to crawlers. His product titles named the part and the brand. The searches that should have landed on exactly the right product landed nowhere.
Bringing fitment ranges and OEM cross-references into the product text, clustering the part-number vocabulary, and controlling which variants entered the sitemap made a very large catalog match a very specific search pattern.