Eighteen years on building sites.
Fourteen years writing code.

I started in construction at twenty. I am thirty-eight. I spent those eighteen years on site, in Luxembourg, learning the trade from the inside: schedules that slip because one trade is waiting on another, quotes redone three times because the client discovers the real scope along the way, jobs that go off the rails not through dishonesty but because nobody had the right information at the right moment.

That is knowledge you cannot acquire by watching the market from outside. You know what a full renovation involves when you have coordinated one. You know why a masonry quote can double when you have priced one yourself.

At twenty-four, I taught myself to code. Alone, no training, without leaving the trade. It was not a career change: it was a second skill, built alongside the first.

Fourteen years on, that is what makes tack.lu possible. I never had to have my knowledge of construction translated by an agency, or explain to a developer why a renovation request involves five trades and not one. Both kinds of knowledge have been in the same head long enough to talk to each other.

Eventually I started my own construction company. And I discovered the problem from the other end: the company looking for work.

A trip out, measuring on site, an evening pricing it up after a full day on the tools — for roughly one quote in five that gets signed. The other four come out of your own pocket, in time and fuel. No small trade business sustains that for long. Decent companies end up not replying to online requests at all, or inflating their prices to absorb the quotes they lose. Either way, the homeowner pays for it.

Like many companies, I paid for the quote platforms on the market. The principle is the same everywhere: a monthly subscription, and in exchange a flow of requests — with no real filter on trade, area or scale. You receive projects that do not match what you do, in municipalities you do not serve, for budgets that bear no relation to your business.

It is not malice: it is what the model produces. When revenue comes from a flat subscription, the quality of the match makes no difference to turnover. Nobody has any incentive to route a request well. And on the other side, the homeowner gets six calls in two days from companies half of which should never have received the request.

I already had both halves of the answer: knowledge of the sector, and the means to act on it. I wrote tack.lu line by line. No fundraising, no sales team, no growth target imposed from outside.

What that changes comes down to one sentence: every decision this platform makes is made by someone who knows what a building site is. Which trades a renovation actually involves. Why a badly qualified request wastes everyone’s time. What a company needs to know before travelling out. And when a homeowner needs a reference price rather than a contact.