Search for a command to run...
At small scale, informal coordination works because everyone can hold the whole company in their head. Past a certain headcount, that stops being physically possible, and structures that felt like bureaucracy before become necessary just to keep decisions moving.
Inflection points show up as symptoms before they show up as headcount milestones: decisions that used to take a day now take two weeks, the same conversation happening in three different meetings, or a founder unexpectedly becoming a bottleneck for approvals that used to be nobody's job to gate.
Functional structures (all engineers together, all marketers together) optimize for depth and consistency but can slow down cross-team execution. Pod or squad structures optimize for speed on a specific outcome but risk duplicated effort and inconsistent standards across pods. Neither is universally correct — the right structure follows from whether your current bottleneck is depth or speed.
A manager with too many direct reports can't give any of them meaningful attention; a manager with too few creates unnecessary layers that slow decisions down. Neither extreme is really about a "right number" — it's about whether the manager can still make good calls fast with the team they have.
Every reorg has a hidden cost: the informal knowledge of who-knows-what-and-why that lived in the old structure. Sequence reorgs to preserve key working relationships where possible, and be explicit about what institutional knowledge needs to be documented before people's roles shift.
Org design isn't a one-time decision made at founding — it's a structure that needs to be revisited deliberately at each inflection point, in response to real coordination symptoms, not just growing headcount for its own sake.