Bounce rate measures single-page sessions divided by total sessions. Someone lands on your page, reads (or doesn't read) it, then closes the tab or hits back. That's a bounce. In Google Analytics 4, this concept shifted to "engagement rate"—the inverse metric—but the principle remains: did the visitor find what they needed or bail immediately? Google has stated publicly that bounce rate isn't a ranking signal. They don't pipe Analytics data into the search algorithm. But dismissing bounce rate entirely misses the point. If 85% of visitors leave your page in three seconds, you likely have a mismatch between search intent and content, slow load times, intrusive interstitials, or misleading title tags. Those issues do affect rankings through user behavior signals Google can measure without Analytics—pogosticking back to the SERP, short dwell time, lack of subsequent searches. Context matters. A local business hours page should have a high bounce rate; users got their answer and left. That's success. An e-commerce category page with 80% bounce rate is a red flag—people should be clicking products. Blog posts sit somewhere in the middle. We see 50–70% as normal for informational content in competitive niches. Under 40% usually means engaged readers exploring related posts. Over 80% consistently means something's broken. At Ottawa SEO, we don't obsess over bounce rate in isolation. We cross-reference it with average session duration, pages per session, and conversion actions in GA4. A high bounce rate with two-minute dwell time on a 1,200-word guide isn't a problem. A high bounce rate with eight-second dwell time on the same guide means rewrite the intro, fix the layout, or re-target the keyword. Improve bounce rate by nailing search intent in the first 150 words, using clear headings and short paragraphs, adding internal links to related content, and ensuring mobile pages load under two seconds. Fix technical friction first—speed, layout shift, pop-ups—then refine the content angle. Bounce rate is a symptom, not the disease.