Carimmat

01The situation

Carimmat matched people by licence plate: you see someone in traffic, you look up the plate, you get in touch. No category, no search demand, and nobody waking up wanting it.

I founded it, ran it, and built the team to five people on €500k raised.

02What changed my mind

Paid acquisition was never going to work: there is no keyword for a product nobody knows exists, and the audience is everyone who drives, which is the same as nobody.

But the idea explains itself in one sentence and journalists like one-sentence stories. The distribution channel was the story, not the budget.

03The decision

Build the product and the press angle together, and treat national coverage as the acquisition channel rather than as a vanity outcome.

04What we shipped

The app on iOS and Android, rebuilt once in 2021, and a launch that landed national television, radio and print, plus a Smart Mobility Award at VivaTech.

05What changed

200,000 registered accounts, from zero, on €500k raised and a five-person team.

I stepped back from day-to-day operations in early 2023 to join lempire. A sale was explored and did not close, and the company was wound down afterwards.

06What I’d do differently

Press gets you the first hundred thousand and none of the next hundred thousand. I built the launch channel well and never built the second one, and a consumer product with no repeat reason to open it cannot survive on coverage alone.

Closing it was the right call and I made it too slowly. Knowing when to stop is a founder skill, and it is the one I was worst at.

I can state what changed and over what period. I would want a clean cohort comparison before claiming the activation rebuild caused all of it, and a longer window to see whether the 40% holds past month one.

The next test is durability. The growth came from a small post-relaunch base, so the real question is whether the positioning keeps working as acquisition scales.

I can state the ARR and the number who switched. I do not have the conversion rate or the cohort behaviour behind them, and I would not claim either without the data.