Loading...
Mira leads product development at a growing startup. When the team was just 4 engineers, their Monday planning meetings took 20 minutes. Everyone spoke, decisions happened fast, and they shipped features weekly. Then the company raised a Series A and hired aggressively. By month three, the engineering team hit 8 people. Mira noticed meetings now ran 55 minutes — not double, but nearly triple. Side conversations broke out. People repeated points others had already made. She calculated: 4 people meant 6 possible conversation pairs. 8 people meant 28. The communication channels hadn't doubled — they'd nearly quintupled. By month six, the team reached 12. Meetings ballooned to 90 minutes. With 66 possible communication pairs, nobody could track who knew what. Mira added a 13th engineer, Kai...
Popular framing: The team is underperforming because of bad processes, poor meeting culture, or insufficient onboarding — all fixable with better management practices.
Structural analysis: Communication channels in a fully-connected team grow as n(n-1)/2, meaning coordination overhead is a nonlinear function of headcount. Past a threshold (~6–8 people), each new hire imposes a net coordination tax on the system that exceeds their marginal output contribution. The team hasn't failed at management — it has exceeded the structural capacity of a single-team topology. The role of 'social proof' in larger meetings—people feel the need to 'perform' or 'reiterate' points to prove their value to the larger group, further bloating the time.
Treating the symptom (long meetings, slow output) as a cultural or process failure leads to interventions — better facilitation, more documentation, targeted hires — that don't address the underlying combinatorial explosion. The real intervention is architectural: decompose the team into smaller autonomous units with minimal synchronization dependencies, reducing the communication graph from O(n²) to O(n).
