← All writing
Engineering04 Oct 20257 min read
The case against microservices for startups under 30 engineers
Daniel Okafor
Principal · AI Systems
Conway's law is real, but so is your runway. A pragmatic look at when monoliths win — and when they don't.
Microservices buy you organisational scaling. They cost you everything else.
Before you have 30 engineers and clear domain ownership, splitting your system is paying a tax for a benefit you don't yet need. Monoliths give you transactions, fewer moving parts, and faster refactors.
The signal that you've outgrown a monolith isn't 'we feel slow'. It's two specific teams stepping on each other in the same module every sprint. When that pattern repeats for a quarter, then carve a service out — not before.
Read next