How to Optimise Backend Performance: A Practical Playbook

TL;DR: Backend performance work is a loop: observe, profile, fix, verify. This post covers the full cycle. Setting up observability, identifying bottlenecks with percentile metrics, applying targeted fixes (N+1 queries, indexing, caching, async offloading), and verifying improvements against p75/p95/p99 latency targets. Why Percentiles Matter More Than Averages Average response time is misleading. An endpoint averaging 80 ms might seem fine until you realise 5% of your users are waiting 800 ms or more. Percentile metrics give you the actual picture: Metric What It Tells You p50 (median) The typical user experience p75 Where the experience starts degrading p95 The worst experience for most users p99 The tail, your worst-case under normal load The goal I worked towards: p95 under 200 ms, p99 under 500 ms, and critical queries completing in under 50 ms. When you optimise, you're compressing the gap between p50 and p99. A fast median with a slow tail means your system is unpredictable, and