Speed gets talked about a lot and measured badly. The page speed statistics below come from large-scale studies — not opinion — so you can benchmark your own site against what the web actually does. These are the page speed statistics worth citing: real load times from millions of pages, conversion data by the second, and how many sites currently pass Google’s own thresholds.
Every number here is attributed to its source and dated. Where a figure is an estimate or comes from a specific vertical, that context is called out, because a stat without a caveat is how bad SEO advice spreads.
How fast the average page actually loads
Backlinko analyzed 5.2 million desktop and mobile pages to benchmark real-world load performance. The headline is that most pages are slower than their owners think. According to Backlinko, the average page is fully loaded in 10.3 seconds on desktop and 27.3 seconds on mobile — and time to first byte alone averages 1.286 seconds on desktop versus 2.594 seconds on mobile.
The mobile gap is not a rounding error. Backlinko found the average web page takes 87.84% longer to load on mobile than on desktop. If you only ever test your site on a fast laptop connection, you are measuring the experience your fewest important visitors get.
- Desktop, fully loaded: 10.3 seconds (Backlinko)
- Mobile, fully loaded: 27.3 seconds (Backlinko)
- Desktop time to first byte: 1.286 seconds (Backlinko)
- Mobile time to first byte: 2.594 seconds (Backlinko)
What actually slows pages down
Weight is the enemy, and the same Backlinko study puts a number on it. Large pages over 3.49 MB take 486% longer to fully load than pages under 0.83 MB, per Backlinko. That is not a gentle penalty for a slightly bigger hero image — it is a five-fold difference driven mostly by images, fonts, and scripts you can trim.
Third-party code is the other quiet tax. Backlinko found that each third-party script added to a page increases load time by 34.1 milliseconds on average. Stack a dozen tags — analytics, chat widgets, ad pixels, A/B tools — and you have added meaningful delay before your own content renders. Backlinko also reported that the fastest JavaScript framework in their sample loaded 213% faster than the slowest, so your stack choice compounds everything else.
Speed and conversions: the money data
Portent’s 2022 research (an update of its earlier 2019 study) tied load time directly to conversion rate across real sites. For B2C e-commerce, the drop-off per second is steep and consistent.
| Page load time | E-commerce conversion rate (Portent, 2022) |
|---|---|
| 1 second | 3.05% |
| 2 seconds | 1.68% |
| 3 seconds | 1.12% |
| 4 seconds | 0.67% |
Read that table twice. According to Portent, conversion rate more than halves between a one-second and a two-second load. Portent also found that an e-commerce conversion rate decreases by an average of 0.3% for every additional second of load time, and that a site loading in 1 second converts roughly 2.5x higher than one loading in 5 seconds.
B2B lead-gen sites show the same shape at a different scale. Portent reported that a site loading in 1 second has a conversion rate about 3x higher than a site loading in 5 seconds, and 5x higher than one loading in 10 seconds. The absolute rates are higher for lead gen, but the direction is identical: every extra second costs you.
Small speed gains, real revenue
The most quoted business case comes from Deloitte’s “Milliseconds Make Millions” study, published in 2020, which measured what happened when mobile site speed improved by just 0.1 seconds across a set of European and US brands. A tenth of a second sounds trivial. The results were not.
- Retail: conversions up 8.4% and average order value up 9.2% (Deloitte, 2020)
- Travel: conversions up 10.1% and average order value up 1.9% (Deloitte, 2020)
- Luxury: page views per session up 8.6% (Deloitte, 2020)
- Lead generation: information-page bounce rate improved 8.3% (Deloitte, 2020)
The point is not that 100 milliseconds is magic. It is that speed improvements compound across every visitor and every funnel step at once, so even small gains move revenue in a way most on-page tweaks never will.
Core Web Vitals: how the web is doing
Google turned speed into a ranking signal through Core Web Vitals, so the fair question is how many sites actually pass. The 2024 Web Almanac, built on real-user Chrome UX Report data, has the answer — and it is humbling. Per the Web Almanac (2024), 43% of websites have good Core Web Vitals on mobile, versus 54% on desktop.
Breaking it down by metric shows where the pain is. According to the 2024 Web Almanac:
- Largest Contentful Paint (LCP): 59% of sites good on mobile, 72% on desktop
- Interaction to Next Paint (INP): 74% good on mobile, 97% on desktop
- Cumulative Layout Shift (CLS): 79% good on mobile, 72% on desktop
LCP — how fast your main content paints — is the metric dragging most mobile sites below the pass line. The Web Almanac also notes that moving from the old FID metric to INP lowered the overall mobile pass rate, meaning real interactivity was worse than the previous metric let on. If under half of mobile sites pass, clearing the bar is a genuine competitive edge, not table stakes.
What this means for your SEO
The data points one direction: most sites are slower than they should be, the biggest losses are on mobile, and small improvements pay back across your whole funnel. Turn the numbers into a short worklist.
- Test on mobile first. With mobile pages loading 87.84% slower on average (Backlinko), your desktop score is flattering you. Benchmark against a real device on a throttled connection.
- Cut page weight before anything else. The 486% penalty on heavy pages (Backlinko) usually traces to unoptimized images and fonts — the cheapest wins on the list.
- Audit third-party scripts. At 34.1 ms each (Backlinko), every tag you can remove or defer buys back render time you are currently giving away.
- Fix LCP first. It is the Core Web Vital failing the most mobile sites (Web Almanac, 2024), so it is where a passing grade is most often won.
This is exactly the kind of diagnosis SEO Rocket is built to run for you. Instead of stitching together a speed tool, a rank tracker, and a spreadsheet, you ask in plain language which pages are slow, which Core Web Vital is failing, and what to fix first — then act on it in the same place. Speed is one input among many, and no single fix guarantees a ranking jump, but the studies above make the trade clear: the sites that get faster convert more and pass thresholds most of their competitors miss.
