Fifteen years in marketing, then I started writing code every day

I did not come to this from a computer science job. I came to it from fifteen years of selling, marketing and publishing other people's products, and then a few years of building my own badly enough to learn how.

The route #

It started at Apple in 2009, running sales training. Then iSquare, the distributor for Greece and Cyprus. Then Antenna, working across the group's properties. Then MAD TV as marketing director, planning across television, radio, digital and events. Then The Flying Pot, where I built the performance marketing and growth unit.

Alongside all of it, for twelve years, I ran TechInGreek — a technology publication that was mine, which meant every decision about it was also mine.

Then a few years of freelance ecommerce work, advising businesses across Europe. And then MyRoomie, where I stopped advising and started building.

What the detour was worth #

The obvious answer is "distribution", and it is true but shallow. Here is the sharper version.

Useful and unknown is the same as not built. Fifteen years of watching good products fail commercially and mediocre ones win burns this in permanently. It is why I think about documentation and positioning while the thing is still being written, not after.

I have watched a lot of people use software badly. Not in usability sessions — in shops, in call centres, in offices where the tool was in the way of the job. That is a different kind of knowledge from reading about UX, and it does not go away.

I can tell the difference between a metric and a result. Performance marketing will teach you this or destroy you. Every number can be moved; most of them mean nothing when they move. Knowing which is which is most of the job, in marketing and in product.

I know what it costs to change someone's mind. Which makes me conservative about products that require it, and much more interested in products that meet people where they already are.

What it cost #

I am not going to pretend the detour was free.

I write code every day now and I am still filling gaps that someone with a conventional route would have filled at twenty-two. Some of them I find by falling into them.

I compensate the way you would expect: I read the source, I write tests, I keep the type checker strict enough to be annoying, and I assume I am wrong for longer than is comfortable. That last one is the marketing background again — fifteen years of being confidently wrong in public will do that.

Where it leaves me #

Founder brain, engineer hands. Neither one is a hobby.

The engineering is real: TypeScript and Next.js at the surface, Go and FastAPI underneath, Tauri when it has to live on a desktop. But the reason I am building housing software rather than developer tools full time is the other half — I spent long enough close to how people actually behave to know which problems are worth the years.