Too many tables? You're probably hurting yourself

A PostgreSQL DB with 10,000+ tables is 40x slower at startup than one with 100 tables. The engine spends time scanning the system catalogs on every connection. It's a real problem, not a corner case — and it shows up when teams normalize data too aggressively.

The fix? Bunch the data into fewer, wider tables. Fewer rows per table doesn't help; fewer tables does. A 100-table schema with 50 columns each loads fast. A 10,000-table schema with 5 columns each crawls.

The rule of thumb: if you can put the data in one table without blowing up, do it. Denormalize early and often. It's the opposite of what your database professor taught you, but it's what the hardware and the query planner actually like.

Why this matters for us: most of the apps primos run on cheap hosting — Heroku free tiers, Render, Vercel — and every extra table costs a few more seconds on startup, which means a slower load for la gente clicking in off WhatsApp. Fewer tables = faster pages for everyone.