Beyond the Code
ES EN

Business

The King of the Castle: Are You the Founder or the Bottleneck?

· by Marc Rubiño

Not long ago I heard Carlos Blanco say something that stuck with me and made me think: "there are two kinds of entrepreneurs: the ones who want to be king and the ones who want to be rich." The rich one knows when to step aside. The king would rather die with his product in his hands than stop controlling everything.

That line hooked me right away, and I couldn't agree more. It made me think about all my years managing teams, building products and helping companies grow: how I've handled it myself, and the kinds of people I've run into along the way, some of whom I've had to deal with.

For example: my first years managing people were at a consultancy in Barcelona, with a team of my own. I had someone with a restless streak, the kind of person who wants to do more than their job description, and who now and then lent another team a hand.

I experienced that as a problem.

I didn't see it for what it was, someone eager to grow who needed a place to do it. I saw a leak. So I closed ranks: I isolated them and made it clear, without ever saying it out loud, that their job was to do what I asked of them.

I did it convinced I was protecting the team. It took me years to understand that I was protecting something else. And nobody told me at the time. Nobody ever does. That's half the problem.

Before I go on, two disclaimers, because this topic gets toxic fast.

This isn't about founders versus professional managers. Both types I'm talking about are founders. Both are all in, both give everything they've got. The difference is smaller than it looks: one builds a company that can run without them and the other doesn't. And one of the two grows more. That's all this article says.

And I say "founder" because that's where it's most visible, but it could be a CEO, a department head, a partner in a firm: anyone can fall into it. In a founder, everything comes together: ownership, identity, credit, fear, twenty years of "I built this by myself". That's why the pattern shows up in its purest form and survives for years without anyone daring to name it.

The king has no ceiling. The king is the ceiling

A company where every decision that matters goes through one person grows at exactly that person's speed. Not an inch faster. You can hire ten people, fifty, two hundred: if the point everything passes through is only so wide, the flow is only so big. Putting more people below a bottleneck doesn't increase the flow. It makes the queue longer.

And I don't mean just their hours. I mean their judgment, their way of selling, their quirks, their blind spots and the three things they've never quite understood. The company inherits the whole package, and whatever they don't master ends up, without anyone ever deciding it, as something that just isn't done around here.

Goldratt laid this out in 1984 for factories, and nobody found it controversial. The only difference between a factory and a company is that a machine can't decide it isn't the bottleneck.

Why you can't see it from the inside

Picture any ordinary afternoon. An order gets complicated, a customer gets nervous, someone on the team is handling it so-so. The boss notices, steps in, picks up the phone and sorts it out in twenty minutes.

And then the worst thing that could happen happens. It works!!

It works fast and it works well. The customer is happy. The team breathes a sigh of relief. And that night the boss goes home with a feeling that happens to be true: if they hadn't stepped in, it would have fallen apart.

What really happened that afternoon is that nobody learned how to solve it. That next time they won't know how either. And that the person who was handling it got a very clear message, even though nobody said it out loud, about what happens when they make the call.

That cost doesn't show up today. It shows up six months from now, disguised as another problem that also has to be solved by stepping in.

Two researchers at MIT, Repenning and Sterman, studied this loop and gave it a name: self-confirming attribution errors. Every time you step in, you get fresh proof that your team can't cope. You never get proof that your team can't cope because you always step in.

That's why telling a king to delegate is useless. His experience proves him right every single day. The mistake is structurally invisible from the inside. You can only see it from the outside, or much later.

From the outside it's obvious, and you can hear it in how they talk

There's a vocabulary. Three things that, if you hear them together and often, tell you exactly what you're dealing with.

"This engineer didn't do what I asked." Never "maybe I didn't explain it well." Never "maybe what I asked for didn't make much sense." The fault always lies downstream, with the last person who touched the thing. This is one of the ones I've run into most. And one of the ones I've said most myself, by the way.

"The business wants this done this way." This is the best one, because it's a small feat of engineering. It turns a personal preference into an institutional mandate you can't push back on, and it conveniently takes the speaker out of the equation. There's no one left to ask why. Next time you hear it, try asking who in the business, in which meeting, and on what grounds. The look on their face is worth more than any report.

And the third isn't a phrase, it's a constant: nothing that goes wrong is ever their fault. Not once. Never.

Put the three together and the math doesn't work: nothing moves around here without my control and, at the same time, nothing that went wrong is my responsibility. Both can't be true at once. Either you decide and you answer for it, or you don't decide and you let go. What doesn't exist is the package where you decide everything and sign off on nothing.

And then there's the line you hear outside the office, over beers after work: "you just can't find good people around here."

Sometimes that's true. But a lot of the time it's simply the explanation that leaves them out of the picture.

The scary version

What I did at the consultancy was the mild form. A first-time manager clinging to what's theirs because they don't know any other way. That can be fixed, if someone tells you or if you happen to think it through.

There's a much worse version, and I've come across that one too.

A person with authority who kept tight control over what other departments were doing: what had to be done, who had to do it and how, while their own department was left to rot. And mind you, it wasn't a lack of time. It was where they'd chosen to spend it. Anything that moved forward without passing through their hands automatically became a problem, and not passively: it got torpedoed, people got turned against it, and the hallways filled with explanations of why it was never going to work.

The striking thing is that from the inside, that doesn't feel like control either. It feels like responsibility. Like being the only one who truly cares whether things turn out well, in a place where everyone else just minds their own business.

And there the ceiling stops being what that person can do. It becomes what that person lets through, which is a lot worse, because it no longer limits just their own area.

I went looking to see if this had a name. It has about half a dozen.

And the odd thing is that none of those bodies of literature cite each other. Each one stumbled onto it on its own and had to christen it without knowing it already had a name next door. Two have already come up here without an introduction: Goldratt's bottleneck and the MIT researchers' capability trap. The other four are these:

Though, to be honest, I still like the king of the castle better than all of them.

That leads to the number that left the worst taste in my mouth. At Microsoft they measured what happened to the ideas teams were sure about before testing them: a third improved the metric, a third did nothing and a third made it worse. A one-in-three hit rate, with whole teams, with data in front of them and money on the line. Now think about the hit rate of one person deciding from memory and in a hurry.

The bill

In case anyone wants it in numbers, it arrives from three directions.

Sure, but founder mode is in fashion now

Founder mode seems to be the coolest thing at the big startups right now, so I know what you're thinking. And you're right to bring it up.

In 2024 Paul Graham published Founder Mode after hearing Brian Chesky explain how the advice to "hire good people and give them room" nearly took Airbnb down. It went viral. And there's data pushing in that direction: Bain analyzed the S&P 500 and founder-led companies delivered returns 3.1 times higher.

That number made me uncomfortable for a while, because it's exactly the opposite of what I'm arguing. I've turned it over in my head and I still haven't fully settled it, but I think the contradiction is only apparent: Bain measures founders who lead, Wasserman measures founders who control. They look a lot alike from the outside and they aren't the same thing.

Graham himself left a line in that essay that his fans quote a lot less: "founders who can't delegate even things they should delegate will use founder mode as an excuse."

And Chesky himself, a year later, unintentionally gave the best summary of all this when asked how he does it: "start in the details, build trust, create muscle memory, and then let go."

That's the whole difference. Not how much you get involved. Whether there's a "and then let go", or whether that line has been twelve years in coming.

Two questions

What stops if you disappear for three weeks without warning? Not a planned vacation with everything tied up neatly. Without warning. If the answer is "everything", what you have is a very expensive job with employees attached.

When was the last time someone told you no and won the argument? If you have to think hard about it, that's your answer: you don't have a team, you have a court. And courts are ruinously expensive, because you're paying salaries to people who know how to decide so that they won't.

For the record, this isn't an article against kings. The founders of Mailchimp never raised a dollar of outside capital in twenty years, and each of them owned 50% the day they sold it to Intuit for twelve billion dollars. Basecamp has spent two decades choosing not to grow, with positive cash flow and a tiny headcount for the revenue it brings in. That's a king, and it's working out great for them, because they chose the castle knowingly and say so out loud.

The problem is the other one. The one who says they want to raise a round and triple in size, and then personally reviews the copy on the website. The one who complains that nobody takes initiative after having patiently, in good faith, built a place where taking initiative comes at a price. The one who's spent three years saying they need a number two and rejecting candidates because none of them would do it exactly the way they would.

That person doesn't have a talent problem in the market. They have a ceiling, and the ceiling is exactly as tall as they are.

And until the day it gets hit, everything looks like it's going reasonably well.

I learned this over time and the hard way, which is how these things get learned. If I had to keep just one of the two questions above, I'd keep the first. It took me quite a few years to be able to answer it without blushing.