Thomas Bue Schultz

This is a holding company’s website, so there is nothing here to buy. No services, no clients, no case studies. Bue Consulting owns an email address I control, and if a message from this domain brought you here to check I am a real person, I am.

The day job

I’m CTO at Adnami, where we build high-impact advertising formats. In practice the work is systems that must answer in the low tens of milliseconds, a few billion times a month, without anyone ever noticing they exist.

Advertising has a reputation and some of it is earned. It is also what pays for most of the free internet: the journalism, the reference material, the tools people use without ever reaching for a card. The realistic alternative is not a cleaner web. It is a smaller one, behind more paywalls, read by fewer people, and I have never found that trade attractive.

So I would rather work on making the mechanism less intrusive and more accountable than stand at a polite distance being disappointed by it. It helps that it is one of the few places left where a slow system cannot hide behind a loading spinner. Either you answered in time or you were not in the auction.

Before that, airports

Seven years at cVation, latterly as VP of Software Development, most of it building software for airports and the aviation industry.

When an airport system fails, the people affected are standing in a building, holding a bag, looking at a screen that is confidently wrong. They cannot close the tab and come back later.

That rewired how I think about failure. Almost no software is genuinely urgent; a little of it really is; and knowing which one you are building ought to change how you build it. Most teams never actually decide. Between airports and advertising the through-line is the same anyway: real-time systems, unglamorous reliability, and teams that have to keep shipping without breaking the floor they are standing on. I have mostly been the person responsible for the floor.

What I keep coming back to

Single authority

One thing owns a given piece of truth and everything else derives from it. Most of the worst bugs I have chased came from two systems each believing it was in charge of the same fact, and each being partly right. It holds for people as well as for data: when two of them own a decision, nobody does.

The right thing at the right time

Not the most, and not the earliest. Building ahead of the need costs about what building behind it costs, and it is much harder to admit to, because while you are doing it it looks like diligence.

Simple where simple is possible

Some problems are irreducibly awkward, and pretending otherwise just moves the complexity somewhere less visible. But a good deal of what gets defended as essential complexity was a choice somebody made once and nobody has revisited since.

Built by a team

Software is a team output, and responsibility sits better on a team than on any individual. One name attached to a system is a risk wearing the costume of accountability: it concentrates the knowledge, it makes one person defensive about their own work, and it quietly excuses everyone else from caring. This sits alongside single authority rather than against it. One place should own a given truth; no single person should own the consequences.

The part I enjoy most is that a team which grows together gets better at all of this in a way that no individual on it could manage alone.

Otherwise

I live in Emdrup, on the north side of Copenhagen, with my girlfriend.

There is also an IMDb page with my name on it, from two Danish films in 1999. I was at the post-production house that did the effects, on the part where footage arrived after the shoot, got digitised, and got recorded back out onto negative. It was the first one-way pipeline I ever worked on, and people remain reliably more impressed by it than by anything I have built since, which I have decided to find funny rather than annoying.