Decisions that seem small at first, but define the future margin. What is the invisible architecture of an early startup?
What most limits a startup is rarely the product.
It is almost always a decision made when everything seemed reversible.
In the first 12 to 18 months, almost everything feels pliable.
The equity can be adjusted.
Roles can be redefined.
The legal form can be ordered below.
The economic model can be “improved when we grow”.
That sense of reversibility is comfortable. And quite misleading.
Not because the founders are improvising. In early stage the focus is where it needs to be: validate product, get customers, survive.
The architecture is for later.
The problem is that some decisions are not tactical. They are decisions that define future headroom. And that is not felt until the company tries to scale.
Property: the cap table that looked temporary
In more than one startup that I accompanied in its first institutional round, the cap table ended up being the main sticking point.
Equity delivered for anxiety.
Advisors with percentages that are difficult to justify.
Founders too diluted too soon.
At the time it helped to move forward.
Years later, it limits options. Not because it is “wrong”, but because it reduces flexibility when you need it most.
In growth, the range of possible decisions matters more than initial comfort.
Engagement: vesting and conversations no one wants to have
Vesting is often treated as a secondary detail. The “we're friends” or “we'll formalize it later” works as long as everything is going well.
When someone leaves, the problem ceases to be emotional and becomes part of the company's model. And the model, when there is already capital and expectations, is not easily rewritten.
It's not about having the perfect contract.
It's about having the right conversation before the contract matters.
Responsibility: roles and decision design
Co-founder roles seem flexible at the beginning. In early stage the ambiguity even helps. Everyone does everything.
I made a mistake there myself.
In one of my companies we distributed roles based on technical capacity. It was obvious to me that a certain person could hold that role. I saw it as simple because for me it was.
We never had the uncomfortable conversation of whether he really wanted to fill that position or if he was comfortable doing it.
It was more of an “ok, I do it” than a conscious decision.
It didn't explode at first. But later we understood that the problem was not one of talent. It was design.
Defining a role is not assigning tasks. It is to understand who decides what, under pressure, with what information and with what real responsibility.
When that's not clear, the startup keeps moving forward. But it does so with invisible friction.
And invisible friction erodes speed.
Legal form: when operational becomes strategic
Legal architecture is often chosen for practicality. Open fast. Operate fast.
It makes sense.
The limit appears when you try to raise international investment or expand cross-border and the legal form does not match the growth model.
What was operational becomes strategic. Late.
Redesigning architecture on the move is almost always more expensive than it seemed in month one.
Economic model: pricing designed to survive
Many founders design prices to close the first contracts. Extensive discounts. Customized conditions. Promises made under pressure.
I myself, in one of my companies, accepted conditions that at the time seemed necessary to close. It worked. But years later I understood the invisible cost of that decision.
The economic model starts to build on exceptions. And exceptions do not scale well.
The model is part of the architecture.
If the model does not support margin growth, neither does the product.
The first question
Today, when I'm presented with a startup or review a pitch, I almost never start with the product.
The first thing I ask is how the partner agreement is defined.
Equity.
Roles.
Vesting.
Who decides what.
The answer to that question tells me much more about the company's architecture than any demo.
Sometimes I discover extremely organized teams.
Other times I realize they never had that conversation.
And that, at an early stage, weighs more than it seems.
None of this usually breaks a company in month six.
Many startups grow the same way.
The problem is that these decisions accumulate. And when growth requires redesigning the architecture or adjusting the model, the cost of change is no longer the same.
Most early founders optimize for speed. It makes sense. Building fast matters.
But some conversations are best had when there is no tension yet.
Not to slow down the onset.
To protect the margin for future decisions.
Because not everything that seems reversible is.
Some decisions simply define the ceiling.
And that is often understood later than one would like.