My query had grown over time, i had added more joins, filters, aggregates and calculations until it became the slowest part of my stack. I tried everything: indexes, denormalization, query hints… nothing. The main Problem: A Bloated Analytical Query On a table with: * 5 million users * 50 million orders * 80 million login events It took over 12 seconds on average. The Fix: Reusable Inline CTEs So, instead of joining raw tables multiple times, I pre-aggregated the heavy tables inside CTEs, so eve