How Fast Should a Website Load for SEO?

The minimum bar to pass Core Web Vitals: mobile LCP under 2.5s, INP under 200ms, CLS under 0.1, at the 75th percentile of real Chrome users over 28 days. To actually win in competitive search, aim for LCP under 1.8s and TTFB under 500ms. Bounce rates climb sharply between 1 and 3 seconds: 53% of mobile users abandon pages that take longer than 3 seconds to load. There is no SEO benefit to loading in 0.5s instead of 1.5s once you are clearly inside the Good band, but conversion and engagement keep improving as you get faster. Benchmarks vary by industry: news sites average around 2.8s LCP, ecommerce around 3.4s, finance around 2.5s.

For SEO, the bar to clear is the Core Web Vitals thresholds on mobile, evaluated at the 75th percentile of real Chrome users over a 28 day window:

• LCP under 2.5 seconds: the largest visible element should render within 2.5 seconds for 75% of visits.
• INP under 200 milliseconds: the worst interaction across the page visit should respond within 200ms for 75% of visits.
• CLS under 0.1: unexpected layout shifts should total less than 0.1 for 75% of visits.

These are the minimum thresholds to pass and qualify for the Page Experience ranking boost. Sites that consistently hit these numbers earn the boost; sites that fail lose it and tend to sit behind faster competitors in close ranking battles.

For the full technical playbook on each metric, see the complete Core Web Vitals guide.

The user behavior data is even more demanding than the SEO bar:

• At 1 second: smooth, no perceived friction. Conversion rates are at the natural maximum.
• At 2 seconds: still inside the Good band, minimal abandonment. Conversion drops 1 to 3% per additional 100ms.
• At 3 seconds: 53% of mobile users abandon. Bounce rates spike sharply. Conversion drops around 20% per additional second.
• At 4 to 5 seconds: catastrophic abandonment on mobile, especially for transactional pages. Users assume the page is broken.
• Above 6 seconds: most mobile users leave before content renders.

Google does not use bounce rate directly in ranking, but the abandonment shows up as pogo sticking back to the search results, which is one of the strongest negative signals a page can produce. For deeper coverage of mobile user behavior, see the mobile website speed guide.

Passing Core Web Vitals is the floor. To consistently win in competitive search results, you want to be clearly faster than the field:

• LCP under 1.8 seconds: comfortably inside the Good band with margin for the 75th percentile.
• INP under 100 milliseconds: well below the threshold, which protects against the regression cycles that ship with new features and plugins.
• TTFB under 500 milliseconds: server response that supports both fast LCP and high Googlebot crawl rates.
• CLS under 0.05: virtually no visible shifting, which signals high engineering quality and protects against future regressions.

Being twice as fast as a passing competitor does not give you twice the ranking boost, but it does give you margin. The 75th percentile rule means your worst 25% of visits matter, and being clearly inside the Good band on your typical visits protects you when devices, networks, or third party scripts get worse than expected.

Average mobile LCP from CrUX field data across major industries in 2026 shows wide variation:

• News and media: median around 2.8s LCP. Heavy ad and video weight makes passing CWV harder.
• Ecommerce: median around 3.4s LCP. Product pages with multiple images, reviews, and recommendation widgets are the usual bottleneck.
• SaaS marketing sites: median around 2.4s LCP. Hero animations and video backgrounds are common problem areas.
• Finance and banking: median around 2.5s LCP. Strong infrastructure helps, but compliance scripts and chat widgets often hurt INP.
• Travel and hospitality: median around 3.2s LCP. Search widgets and image heavy galleries push LCP up.
• Healthcare: median around 3.0s LCP. Marketing scripts and accessibility widgets are common bottlenecks.
• Government: median around 4.0s LCP. Legacy infrastructure and minimal optimization budget show in the data.

The practical takeaway: in any industry, being meaningfully faster than the median is enough to win the speed tiebreaker against most competitors. You do not need to be the fastest site in the world; you need to be faster than the pages you are competing with.

Speed targets vary by device and connection type:

• Mobile (4G, mid range Android): this is the primary SEO target. LCP under 2.5s, INP under 200ms, CLS under 0.1. Most sites that fail CWV fail here.
• Mobile (5G, modern flagship): targets are easier to hit but should still be the same. CrUX reports every visit, including the slower ones.
• Desktop (cable, modern laptop): LCP under 1.8s and TTFB under 300ms are realistic targets. CrUX reports desktop separately and Google evaluates both.
• Tablet: aggregated with mobile in most reports. Target the mobile thresholds.

The mobile-desktop gap is not a rounding error: our study of the top 500 global sites found 84% of them load slower on mobile than desktop and only 44% pass Core Web Vitals on mobile versus 64% on desktop. Do not optimize only for the device class you use yourself. Test on mid range Android over 4G to see what your real users experience. Most performance failures happen in this profile, not on the developer's MacBook.

For large sites (10,000+ URLs), speed also affects how much of your site Googlebot can crawl:

• TTFB under 200ms: Googlebot crawls at the maximum rate the host can support.
• TTFB 200 to 600ms: crawl rate is moderate. Most sites sit here.
• TTFB 600ms to 1s: Googlebot throttles back to avoid overloading the server. New and updated pages get indexed more slowly.
• TTFB above 1s: Googlebot reduces crawl significantly. Large sites lose indexation coverage for their long tail.

For a 100,000 URL site, the difference between 200ms TTFB and 800ms TTFB can mean tens of thousands of pages per month in crawl coverage. If you publish frequently or run a large content site, server speed has outsized SEO value beyond the direct CWV signal.

There is a diminishing returns curve for SEO. Once you are clearly inside the Good band on all three Core Web Vitals, additional speed work does not add direct ranking benefit:

• A page with 2.0s LCP and a page with 1.0s LCP both pass and both qualify for the Page Experience boost. Google does not award extra ranking for being twice as fast.
• Once CWV pass, additional speed work pays back through conversion lift and engagement, not through additional ranking gains.
• The exception is when your competitors are also passing CWV. Then being meaningfully faster gives you margin against future regressions and slightly better engagement signals.

The right SEO investment shape: get to passing CWV first, then maintain that status while you put marginal effort into content, intent, and link work. Once you can no longer find a missing topic or a low quality competitor to outrank, come back to speed and push deeper.

If you need help getting from failing to passing Core Web Vitals, our Core Web Vitals optimization service owns the diagnosis and the fixes.

Metric	Good	Needs Improvement	Poor
LCP (Largest Contentful Paint)	under 2.5s	2.5 to 4.0s	over 4.0s
INP (Interaction to Next Paint)	under 200ms	200 to 500ms	over 500ms
CLS (Cumulative Layout Shift)	under 0.1	0.1 to 0.25	over 0.25
TTFB (Time to First Byte)	under 0.8s	0.8 to 1.8s	over 1.8s
Full page load (total)	under 3s	3 to 5s	over 5s