Bounce rate is a context metric, not a quality metric
A 90% bounce rate sounds catastrophic until you realize the visitor read your full article and left satisfied. Single-page landing pages with one CTA have ~90% bounce by definition — Google Analytics 4 changed the bounce definition to address this (now requires <10s engagement to count), but legacy tools still mismeasure.
How to actually use bounce rate
Compare your bounce rate against your own historical baseline for the same traffic source. A spike on paid traffic usually means message mismatch (the ad promised X, the page delivers Y). A drop usually means improved relevance or speed. Bouncing at the same rate as your competitors is meaningless — what matters is whether you're improving.
Reading bounce + engagement together
High bounce + low engagement = bad page. High bounce + high engagement = visitors read what they came for and left (often fine). Low bounce + low engagement = visitors clicked off-CTA links and stuck around without converting (usually a navigation problem). Run a full analysis to find what's actually causing your bounces.