The State of Web Vitals

LCP distribution

How LCP is distributed across real-user data, plus pass-rate breakdown.

Field data PhoneDesktopAll Scope All sites Q2 2026 edition · Desktop field outcomes
1

At a glance the headline numbers for LCP distribution

How LCP is distributed across real-user data, plus pass-rate breakdown.

86.0%
of sites pass LCP
1.4s
Typical site
2.8s
Worst 10% of sites

86.0% of sites pass LCP. The typical site's LCP is 1.4s. The worst 10% are above 2.8s.

The State of Web Vitals · Q2 2026 · 189,915 sites · desktop field datacorewebvitals.io/state-of-cwv
2

The LCP distribution site count at each LCP, from good to poor

0 2,442 4,883 7,325 9,766
0ms–250msms41 sites0.1% of all sites
250ms–500msms1,535 sites2.5% of all sites
500ms–750msms5,741 sites9.3% of all sites
750ms–1sms9,246 sites14.9% of all sites
1s–1.3sms9,766 sites15.8% of all sites
1.3s–1.5sms8,564 sites13.8% of all sites
1.5s–1.8sms6,797 sites11% of all sites
1.8s–2sms5,064 sites8.2% of all sites
2s–2.3sms3,860 sites6.2% of all sites
2.3s–2.5sms2,716 sites4.4% of all sites
2.5s–2.8sms2,067 sites3.3% of all sites
2.8s–3sms1,491 sites2.4% of all sites
3s–3.3sms1,152 sites1.9% of all sites
3.3s–3.5sms861 sites1.4% of all sites
3.5s–3.8sms625 sites1% of all sites
3.8s–4sms492 sites0.8% of all sites
4s–4.3sms383 sites0.6% of all sites
4.3s–4.5sms302 sites0.5% of all sites
4.5s–4.8sms231 sites0.4% of all sites
4.8s–5sms189 sites0.3% of all sites
5s–5.3sms143 sites0.2% of all sites
5.3s–5.5sms131 sites0.2% of all sites
5.5s–5.8sms95 sites0.2% of all sites
5.8s–6sms67 sites0.1% of all sites
6s–6.3sms55 sites0.1% of all sites
6.3s–6.5sms56 sites0.1% of all sites
6.5s–6.8sms48 sites0.1% of all sites
6.8s–7sms31 sites0.1% of all sites
7s–7.3sms40 sites0.1% of all sites
7.3s–7.5sms23 sites<0.1% of all sites
7.5s–7.8sms22 sites<0.1% of all sites
7.8s–8sms22 sites<0.1% of all sites
8sms and up135 sites0.2% of all sites
p50 = 1.4s
p75 = 2s
p90 = 2.8s
p99 = 5.4s
0ms–250ms 250ms–500ms 500ms–750ms 750ms–1s 1s–1.3s 1.3s–1.5s 1.5s–1.8s 1.8s–2s 2s–2.3s 2.3s–2.5s 2.5s–2.8s 2.8s–3s 3s–3.3s 3.3s–3.5s 3.5s–3.8s 3.8s–4s 4s–4.3s 4.3s–4.5s 4.5s–4.8s 4.8s–5s 5s–5.3s 5.3s–5.5s 5.5s–5.8s 5.8s–6s 6s–6.3s 6.3s–6.5s 6.5s–6.8s 6.8s–7s 7s–7.3s 7.3s–7.5s 7.5s–7.8s 7.8s–8s 8s+
Good (≤2.5s) Needs improvement Poor (>4s) Percentile markers Total: 61,991 sites

Most sites pass LCP. 86.0% sit at or under 2.5s. The typical site is at 1.4s, and 75% of sites stay under 2.0s. Only 14.0% fail LCP. The slowest 10% are above 2.8s, and the slowest 1% are above 5.4s.

The State of Web Vitals · Q2 2026 · 189,915 sites · desktop field datacorewebvitals.io/state-of-cwv
3

The LCP pass rate the share of sites that are good, needs improvement and poor

LCP
86%
10.8%
3.2%
Good Needs Improvement Poor

86.0% of sites pass LCP. 10.8% are in the needs improvement band, between 2.5s and 4s. 3.2% are poor, above 4s. A miss is usually needs improvement, not poor.

The State of Web Vitals · Q2 2026 · 189,915 sites · desktop field datacorewebvitals.io/state-of-cwv
4

Why this matters for the Core Web Vitals, and where to start fixing it

Largest Contentful Paint is when the biggest thing in the viewport finishes rendering, usually the hero image or a headline. It is the moment the page stops looking blank and starts looking loaded, so it is the metric your visitors feel first. A slow LCP is almost always one of a few things: the server takes too long to respond, the image is too big or discovered too late, or render-blocking CSS and fonts hold everything up.

Start with what paints. Find the LCP element, make sure the browser can discover it early, and give it a clear path to the screen. Preload the hero image, serve it in a modern format at the right size, and keep render-blocking resources out of the way. The server side matters too, but the biggest wins are usually in how the main image and the critical CSS are delivered.

How are sites doing on LCP?

86.0% of sites have a good LCP. The typical site sits at 2.0s at the 75th percentile; the slowest 1% pass 5.4s.

Go fix it Preload the LCP image → Fix slow hero images → Optimize images for Core Web Vitals → Resource prioritization guide →
Related signals TTFB → FCP → CLS → INP → Chrome field data from 189,915 sites, representing millions of real page loads · How we measured
Live queries (2) — admin only
Query #1: 60.8 ms 
SELECT COUNT(*) AS count,
            COUNT(*) FILTER (WHERE crux."desktop".lcp <= 2500) * 1.0 / COUNT(*) AS good,
            COUNT(*) FILTER (WHERE crux."desktop".lcp > 2500 AND crux."desktop".lcp <= 4000) * 1.0 / COUNT(*) AS needs_improvement,
            COUNT(*) FILTER (WHERE crux."desktop".lcp > 4000) * 1.0 / COUNT(*) AS poor,
            quantile_disc(crux."desktop".lcp, 0.10) AS p10,
            quantile_disc(crux."desktop".lcp, 0.25) AS p25,
            quantile_disc(crux."desktop".lcp, 0.50) AS p50,
            quantile_disc(crux."desktop".lcp, 0.75) AS p75,
            quantile_disc(crux."desktop".lcp, 0.90) AS p90,
            quantile_disc(crux."desktop".lcp, 0.99) AS p99
            FROM sites WHERE 1=1 AND crux."desktop".lcp IS NOT NULL;
Query #2: 64.6 ms 
SELECT CASE WHEN crux."desktop".lcp >= 0 AND crux."desktop".lcp < 250 THEN 0 WHEN crux."desktop".lcp >= 250 AND crux."desktop".lcp < 500 THEN 1 WHEN crux."desktop".lcp >= 500 AND crux."desktop".lcp < 750 THEN 2 WHEN crux."desktop".lcp >= 750 AND crux."desktop".lcp < 1000 THEN 3 WHEN crux."desktop".lcp >= 1000 AND crux."desktop".lcp < 1250 THEN 4 WHEN crux."desktop".lcp >= 1250 AND crux."desktop".lcp < 1500 THEN 5 WHEN crux."desktop".lcp >= 1500 AND crux."desktop".lcp < 1750 THEN 6 WHEN crux."desktop".lcp >= 1750 AND crux."desktop".lcp < 2000 THEN 7 WHEN crux."desktop".lcp >= 2000 AND crux."desktop".lcp < 2250 THEN 8 WHEN crux."desktop".lcp >= 2250 AND crux."desktop".lcp < 2500 THEN 9 WHEN crux."desktop".lcp >= 2500 AND crux."desktop".lcp < 2750 THEN 10 WHEN crux."desktop".lcp >= 2750 AND crux."desktop".lcp < 3000 THEN 11 WHEN crux."desktop".lcp >= 3000 AND crux."desktop".lcp < 3250 THEN 12 WHEN crux."desktop".lcp >= 3250 AND crux."desktop".lcp < 3500 THEN 13 WHEN crux."desktop".lcp >= 3500 AND crux."desktop".lcp < 3750 THEN 14 WHEN crux."desktop".lcp >= 3750 AND crux."desktop".lcp < 4000 THEN 15 WHEN crux."desktop".lcp >= 4000 AND crux."desktop".lcp < 4250 THEN 16 WHEN crux."desktop".lcp >= 4250 AND crux."desktop".lcp < 4500 THEN 17 WHEN crux."desktop".lcp >= 4500 AND crux."desktop".lcp < 4750 THEN 18 WHEN crux."desktop".lcp >= 4750 AND crux."desktop".lcp < 5000 THEN 19 WHEN crux."desktop".lcp >= 5000 AND crux."desktop".lcp < 5250 THEN 20 WHEN crux."desktop".lcp >= 5250 AND crux."desktop".lcp < 5500 THEN 21 WHEN crux."desktop".lcp >= 5500 AND crux."desktop".lcp < 5750 THEN 22 WHEN crux."desktop".lcp >= 5750 AND crux."desktop".lcp < 6000 THEN 23 WHEN crux."desktop".lcp >= 6000 AND crux."desktop".lcp < 6250 THEN 24 WHEN crux."desktop".lcp >= 6250 AND crux."desktop".lcp < 6500 THEN 25 WHEN crux."desktop".lcp >= 6500 AND crux."desktop".lcp < 6750 THEN 26 WHEN crux."desktop".lcp >= 6750 AND crux."desktop".lcp < 7000 THEN 27 WHEN crux."desktop".lcp >= 7000 AND crux."desktop".lcp < 7250 THEN 28 WHEN crux."desktop".lcp >= 7250 AND crux."desktop".lcp < 7500 THEN 29 WHEN crux."desktop".lcp >= 7500 AND crux."desktop".lcp < 7750 THEN 30 WHEN crux."desktop".lcp >= 7750 AND crux."desktop".lcp < 8000 THEN 31 WHEN crux."desktop".lcp >= 8000 THEN 32 END AS bucket_idx, COUNT(*) AS count FROM sites WHERE 1=1 AND crux."desktop".lcp IS NOT NULL GROUP BY bucket_idx;
JSON file lookups (4)
20260625/meta.json 0.03 ms
20260625/types.json 0.65 ms
20260625/loaf-scripts.json 0.29 ms
20260625/menu.json 0.18 ms