The instinctive response to growth is to add people. More demand, more headcount; more complexity, more teams. It feels like scaling. Often it is the opposite — adding mass to a system whose real constraint was never the number of hands, and watching the new people make the same unscalable decisions faster and in more places.
What actually scales, or fails to, is decisions. A growing organization is a decision-making machine, and its limit is reached when too many decisions require the same few people, or when the same decision gets made differently depending on who happens to make it. Adding people to that does not relieve the constraint; it multiplies it. You do not scale by growing the team. You scale by reducing the number of decisions that require a human at all — by encoding judgment into structure so the right outcome is the default, not the result of someone good being in the room.
This is the real cost of hero dependency. The brilliant engineer who holds the system in their head, the founder who personally approves the important calls — at scale they are not assets; they are single points of failure the org chart is too polite to label. A great team can mask a missing structure for a surprisingly long time, which is exactly the danger: the masking ends suddenly, usually at the worst moment, and looks like a people problem when it was a design problem all along.
So the question under every reorg is not "do we have enough people." It is "which decisions still depend on specific humans, and which of those should be structure instead." Teams don't break systems. Systems — the ones that never turned judgment into design — break teams, one heroic exception at a time.