A scale-up is a company that has proved the thing works, with established product-market fit and compounding revenue, but has not yet proved it can run as an institution rather than as a founder's extended hand; and the long, uneven transition between those two modes is where most growing companies break.
The first mode is the one founders mean when they talk about their startup days, founder-led and fast, with coordination happening because everyone knows everyone and shares a single mental model of what the company is and how it works. The founder makes decisions, alone or in conversation with whoever is around, and information moves through hallways and Slack and lunch.
The second mode is what mature companies look like, where the coordination work that used to be done by everyone-knows-everyone is done instead by structure: decision rights named explicitly so that people do not have to guess, roles scoped tightly enough that someone who has not absorbed the founder's mental model can still do the job, channels for information that do not depend on people sharing a hallway. What this buys is the ability to onboard senior people into specific jobs and have them produce results without first having to learn the founder.
The journey between these two modes is not a single decision or a single project but a multi-quarter, often multi-year unfolding, and three of its properties make it unusually hard.
The first is that the journey is continuous rather than discrete. There is no day on which the company stops being a startup and starts being an institution; instead, you build institutional infrastructure while still relying on startup behaviours, and the gap between the two is where the friction lives. Different parts of the company end up expecting different operating modes simultaneously, sometimes within the same meeting, and the cognitive dissonance lives mostly inside the founder, who is trying to be both kinds of leader at once.
The second is that the journey is path-dependent: what you build at thirty people becomes the constraint at eighty. A workflow that someone sketched on a Friday afternoon in year two because something needed to ship will still be operating as law in year four, by which point it is bottlenecking the entire revenue organisation. An early hire brought in for a temporary problem becomes a VP of a function the company no longer needs, and you cannot gracefully remove them because the company has built around them. Decisions made under time pressure outlive their usefulness by years, and the cost of that lag is paid quarterly.
The third is that the founder has to redesign their own role, and no other executive transition asks the leader to consciously dismantle the way they personally operate. A new public-company CEO inherits a structure and either keeps it or alters it, but their identity is not fused with how the company runs; a founder at a scale-up is the structure, in the sense that decisions route through them, the standards live in their head, and the culture is whatever they pay attention to. Redesigning the structure therefore means redesigning what the founder personally does, which is psychologically much more brutal than the management literature acknowledges.
The work in scale-up land is therefore structural rather than operational, in the sense that most of the operating problems consuming founder attention are downstream of structural choices that have stopped fitting. Those choices were right at the time they were made, which is why they survived long enough to become structural in the first place. Recognising that, and naming the specific choices in your company that no longer fit, is where the work begins.