Guessing at performance problems wastes time. The Profiler can be used to find what is truly slow before optimizing anything.

Code splitting reduces initial load time, but only when applied at the right boundaries. Here is what tends to work in practice.

Slow list rendering has a small number of truly common causes. Here is the diagnostic process I really use, with the specific fix for each cause.

Context is undeniably convenient but causes specific, measurable performance issues when misused. Here is what tends to show up when profiling context-heavy applications.

Bundle size affects load time directly, but you need to measure what is truly in your bundle before you can meaningfully reduce it. Here is a process for doing so.

Understanding how React batches state updates explains a lot of confusing render behavior. Here is what is happening under the hood, based on testing across different scenarios.

This choice is often framed as a performance decision, but the measured difference is smaller than commonly assumed. Here is what testing showed in practice.

Key prop mistakes cause subtle performance and correctness issues that are easy to miss. Here is what to watch for when debugging these specific issues.

These two techniques are often confused, and getting the implementation wrong defeats the actual performance benefit. Here is what distinguishes the two approaches.

Seeing double renders in development with Strict Mode confuses many developers into thinking something is broken. Here is what is really happening and why.