Why founders end up trapped by the leadership team they built

Ask a founder who has just closed their Series A to describe the leadership team and watch what they reach for. They point at the org chart. Head of Engineering here, a commercial lead finally hired, operations and finance each with a capable person in the box at last. After years of carrying those functions alone, the founder looks at the chart and feels something settle. The team is built.

Well, the chart might be complete. But the team does not exist – at least not yet.

More than thirty years ago, Jon Katzenbach and Douglas Smith drew a distinction every scaling founder should study: the difference between a working group and a real team. A working group is a set of strong individuals who each deliver in their own lane and meet mainly to share information; their output is the sum of individual bests. A real team produces something none of them could alone and holds itself mutually accountable for the result. It meets to make decisions. The two look identical on a chart and behave nothing alike.

Most founders, through the first institutional rounds, build the first and name it the second, in a structure with only one real point of connection: themselves. Each function reports to the founder, spoke to hub. So, every key decision that touches two functions at once has one place to go, up the spokes and back down through the one person connected to everyone. The founder has become the integration point for the whole company. Early on, that is an asset. It is about to become the thing that holds both the company and the founder back.


A chart is not a team

The hub-and-spoke is survivable while the company is small. It stops being survivable at a certain threshold, and in DeepTech that threshold is identifiable. On the best current data, a DeepTech business stops being an R&D effort and becomes a commercial organisation somewhere in the 51-to-200 employee range, according to this year's global study of more than 13,000 DeepTech startups from HTGF and TUM. That is exactly when functional heads must start behaving like a team, because the problems stop fitting inside single functions.

A working group can ship a product. It struggles to make the calls that decide whether a company scales, because those calls live between functions, not inside them. When engineering, commercial and finance must resolve a significant trade-off together, a working group refers it upward; a real team resolves it in the room. The distinction is invisible on the chart and decisive in practice. Seeing capable people deliver in their own lanes, the founder concludes the team is working, when it has not yet been tested on the one thing a team is for: deciding together.


Why the technical founder builds spokes

DeepTech founders build the hub-and-spoke model without noticing. The founder is usually the person whose name is on the science, and their deepest working relationships are the vertical ones: to each technical problem, to each expert hire. What they are not building, because they have never had to, is the machinery that lets those hires work across each other without the founder in the middle.

This is not about whether senior people get along; mature executives manage that themselves. It is about the beginnings of an operating system: the shared way decisions get made, recorded and repeated, which a large company institutionalises and takes for granted, and a young company does not have unless someone builds it. In a hub-and-spoke, the founder is that operating system. Routing, memory and arbitration all run through one head, because no system yet does.

And on each connection, at least through the early stages, the founder genuinely is the most capable person in the conversation. That is simply the startup mechanism. The same study found that the value of a founder's engineering experience peaks early, at around five years, while the value of management experience keeps compounding for far longer. The technical mastery that built the company stops being the thing that scales it, well before the founder stops relying on it. And because the founder is still right often enough, no one else is ever forced to become the person who decides.


The trap the science sets

In a fast-feedback business like SaaS, the market breaks this pattern early. DeepTech grants no such mercy: the development cycle may run for years, so the founder-as-builder identity continues simply as the default.

That is the trap. In SaaS, the pace forces the founder out of the centre within a year or so. In DeepTech nothing forces the issue, so the founder can remain the integration point for years, reading their own continued indispensability as evidence that the structure is sound.

And it is worth highlighting who this catches. Not the weak founder, whose company rarely survives long enough or scales far enough to develop the problem at all. This is a trap you earn by being good. The founder capable of building something that reaches Series A and B is precisely the one who can hold the centre convincingly for years, which is exactly why the warning never comes.


The backdrop that makes it structural

Meanwhile the operating environment keeps raising the stakes. Each round brings larger, more expectant investors, increasingly from overseas in European DeepTech, because the domestic market thins precisely at the growth stage. Europe converts only about 10% of DeepTech startups from Series A to Series B, against roughly 24% in North America, and more than half of later-stage capital comes from outside the region. Each round expects a more autonomous company than the last.

But here is what the scaleup founder faces that the established company does not. A mature company runs on decision systems laid down over years, and they hold whoever occupies the seats: a functional head can leave, and decisions still get made, because how they get made does not live in any one person's head. The scaleup has no such inheritance. Its operating system is still forming, reconfigured at every round, hopefully becoming less and less reliant on the founder's connective tissue. The mature company's advantage is not that its leaders know each other better. It is that the company keeps working when they change.


The tell

You can diagnose a hub-and-spoke from the inside, if you know the symptom. Two of your senior people cannot resolve a question that sits between their functions without one of them coming to you. Not because they are weak, and not because the question is hard, but because they have no direct route to settle it between themselves. The only path runs through you.

Most founders read this as being needed and take quiet satisfaction in it. It is the opposite: the signal that the team has never once had to function as one. And it carries a daily cost you rarely name correctly. Every cross-functional decision with a founder-shaped hole in it is an hour spent as the switchboard operator rather than on the work only you can do.


The move that looks like delegation, and isn't

So, the founder, seeing the bottleneck at last, resolves to delegate. And here most make the same move, and it fails the same way. They hand over tasks, name owners, ask for outcomes, step back from the doing. Then they wait to feel freed, and do not, because the bottleneck has not moved.

They delegated the work and kept the decisions. Tasks were handed out; the authority to decide stayed at the hub. A team that can execute what you assign but must come back the moment a situation falls outside the brief has not been delegated to. It has been handed a to-do list. The work is distributed and the judgement is not, which is the hub-and-spoke again, wearing the costume of delegation.

Real delegation hands over the decision, not just the task, and here the earlier distinction earns its keep. The path from a working group to a real team, Katzenbach and Smith observed, runs through conflict: people become a team by working through the hard trade-offs together. A founder who keeps every decision at the hub ensures that conflict never happens, and so ensures the team never forms. Push the decisions out to where the work is, and you are not abdicating; you are letting a real team come into being.

The decisions worth keeping are the few only the founder can make. The rest you hold on to because you still make them faster and better than anyone in the building, which feels like efficiency but works like a brake. The skill is telling the two apart, and the ones truly yours to keep will be far fewer than initially feels comfortable.

Because the decisions you free yourself from are what buy your own transition. Every call that resolves without you is time returned to the work with no one else's name on it: the strategy development, the capital raising, the long arc of where the company is going. The founder trapped as the integration point cannot become the CEO the next round is looking for, not for want of capability, but because the wiring keeps them in the switchboard. Building a real team is how a founder is finally freed to lead one.


What to ask

None of this shows up on the chart, so stop looking at it. When two functions disagree, does it reach you before it reaches a resolution? When you take a week away, does the company decide or store the decisions up for your return? Are your senior people accountable to each other for what you achieve together, or each to you for their own lane alone?

If the honest answers point back at you, you have not built a team. You have built a group and made yourself the only thing connecting it together. The work now is not to hire another impressive individual. It is to hand the decisions to the people you already have, and to discover, in the space that opens, the job you were always supposed to be doing.

 
This is the first of three notes on building a leadership team that runs without you at its centre. The next looks at the room where a team is either forged or quietly kept a group: the leadership meeting, and what it is actually for. The third turns to the hardest part, getting the right senior people into the right seats, and what it takes from the people who got you here.

Let's talk.