Most SaaS products don't fail because of bad code. They fail because the founders optimized for the wrong constraints — usually the ones loudest on Twitter.
Scale is a side effect of three boring decisions: a clean data model, idempotent background jobs, and a pricing page that doesn't apologize. Get those right and most other problems become tractable.
The single biggest lever I've seen across 200+ products is queueing. Move anything that isn't a direct user response into a job. Your p95 will drop, your error budget will recover, and your on-call will stop dreading Sundays.
Don't build for a million users on day one. Build for ten thousand who love it — the architecture for the next zero will reveal itself.
