Bounce Rate
What drives users away in the first ten seconds — benchmarks, causes, and the fixes that actually move the number.
People decide to stay or leave in 3–7 seconds. The decision turns on perceived load speed, clarity of value prop, and trust signals — in that order. Optimize those three before chasing anything else.
Realistic benchmarks by page type
"Good bounce rate" depends on the page's job. Don't compare a blog post to a checkout page.
Causes and fixes
What we have actually moved in our own data and in customers' sites.
Slow load time (LCP > 2.5s)
Fix: Compress images, defer non-critical JS, edge cache. 1 second of LCP delay = ~7% conversion loss.
Unclear value prop above the fold
Fix: In 5 seconds, visitor must know what you do + who it's for. Specific beats clever.
Aggressive popups / cookie bars
Fix: Defer popups 15+ seconds. Cookie bars at bottom, not blocking content.
No social proof in viewport
Fix: Logos, testimonials, star ratings near the hero. Trust signals reduce immediate bounce.
Too many CTAs
Fix: One primary action per page. Visitors who can't decide leave.
Layout shift (CLS > 0.1)
Fix: Reserve space for images and embeds. Fonts loaded with font-display: swap and explicit sizes.
What we changed and what happened
Hero copy: "The all-in-one platform for modern businesses."
Hero copy: "Form your LLC, run payroll, and pay your taxes — in one place."
Bounce rate fell ~12 pts on /home
Cookie banner blocking the hero (modal style)
Cookie banner at the bottom, 60% opacity, dismissable
Bounce rate fell ~6 pts; cookie consent rate stayed flat
Three CTAs above the fold: "Start free trial", "Watch demo", "Talk to sales"
One CTA: "Start free trial" (secondary links in subnav)
Sign-up rate from /home rose ~18%, bounce flat
Hero image: stock photo of a team in an office
Animated product screenshot looping in 8 seconds
Time-on-page rose ~24 seconds; bounce fell ~4 pts
Common diagnostic mistakes
Looking at site-wide bounce rate.
Always segment by page + traffic source. A high site-wide number can hide one bad page or one bad campaign.
Comparing your bounce rate to a competitor's.
You don't have their data. Compare to your own historical baseline and to your specific page type.
Trying to fix bounce with more CTAs.
More CTAs increase bounce, not decrease. Decision fatigue. One clear CTA wins.
Optimizing Lighthouse score instead of real-user metrics.
CrUX (Chrome User Experience Report) data is what Google ranks on. Synthetic Lighthouse runs are a starting point, not the goal.
Triage checklist when bounce spikes
- Check page speed (LCP) — did a recent deploy regress it?
- Check traffic source breakdown — did a campaign change targeting?
- Check device split — mobile-only regression suggests responsive break
- Pull session recordings (Hotjar / FullStory) for 20 bounced sessions
- Compare scroll depth on bounced vs non-bounced — where do bouncers stop?
- Look at search query data — are visitors landing on wrong page for their intent?