Brands didn't want traffic. They wanted customers.
How Carmate went from a new-car catalogue to a lead-generation business, and the decisions along the way.
The market turned to new cars
In 2024 it became clear that Cambodia's car market was moving from used cars to new ones. Carmate had been built for used cars: a simple listing, posted and browsed. New cars needed something else entirely: a catalogue of every official brand, model, year and trim, with prices people could trust.
I designed that catalogue from the data structure up. It's a story of its own.
Complete, but just a catalogue
Local staff filled in the catalogue while I built the website around it. When it went live, two things were wrong:
Give each side what it needs
Buyers in Cambodia want to see real cars in stock and the current promotions. Brands need proof that people are coming. So I designed Seller Central as an MVP: brands post in-stock cars and promotions, and see daily visitors, impressions, views and clicks on every button across their models, listings and promotions.
A great response, and the wrong question
From mid-March 2026 I took a presentation to several official brands. The response was better than I expected:
Then two problems. They measured our traffic against the couple of hundred thousand people their Facebook ads reach. And every meeting came back to one question: “So how many customers can you bring us?”
“Brands don't buy traffic. They buy customers.”
I had seen this before. Used-car dealers who ignored our calls became very responsive the moment we brought a buyer to their shop. Official brands were no different.
Turn the catalogue into a lead engine
In April 2026 I started the redesign. Check Offers and test-drive forms went onto every model, listing and promotion. Seller Central was rebuilt around leads. Traffic charts moved to the back, and the monthly brand report now shows leads and which parts of the site brought them in.
Getting that redesign launched is case study 02.
A smaller number I could stand behind
To show which part of the site triggered each enquiry, I started tracking every element. That exposed scrapers: automated visits inflating our numbers. I took them out of every report, even though our figures got smaller.
A report that holds up when a brand checks it is worth more than a bigger one that doesn't.
Real demand, and the next problem
No brand has signed yet. The lead data is what I'm taking back to them, as a month-by-month story of what Carmate sends each brand.
The next problem is already clear: 1,049 people pressed Check Offers and 41 finished the form. I've added tracking to find where the rest drop off.
What I'd do differently
I'd put a lead form in front of brands before building traffic dashboards. “How many customers can you bring?” was the real question all along. Asking it in the first meeting would have saved a round of building.