Connection hints

dns-prefetch, preconnect and preload links the document declares.

Field data PhoneDesktopAll Scope All sites Q2 2026 edition · Phone field outcomes
Metric LCP INP CLS
1

At a glance the headline numbers for Connection hints

dns-prefetch, preconnect and preload links the document declares.

0
dns-prefetch hints
on the typical page
0
preconnect hints
on the typical page
0
preloads
on the typical page
180,899
sites measured
phone field data

Connection hints on the typical site: 0 dns-prefetch, 0 preconnect, 0 preload.

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

Distribution & median LCP site count and median LCP at each level of connection hints — dns prefetch

0ms 750ms 1500ms 2250ms 3000ms
2.5s
0 61448 122895
0–0 1–1 2+
Good (≤2.5s) Needs improvement Poor (>4s) Site count
The State of Web Vitals · Q2 2026 · 189,915 sites · phone field datacorewebvitals.io/state-of-cwv
3

Passing LCP by connection hints — dns prefetch which level passes the LCP most often

Connection hints — dns prefetchSitesPassing LCPLCP
0–0 122,895 85% 1.5s
1–1 27,829 75% 1.8s
2+ 30,175 74% 1.8s
Good Needs Improvement Poor Faded rows: under 100 sites

Connection hints — dns prefetch 0. p75 1. p99 8. At the low end (0–0): LCP 1.5s. At the high end (2+): LCP 1.8s. computed

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

Distribution & median LCP site count and median LCP at each level of connection hints — preconnect

0ms 750ms 1500ms 2250ms 3000ms
2.5s
0 64859 129717
0–0 1–1 2+
Good (≤2.5s) Needs improvement Poor (>4s) Site count
The State of Web Vitals · Q2 2026 · 189,915 sites · phone field datacorewebvitals.io/state-of-cwv
5

Passing LCP by connection hints — preconnect which level passes the LCP most often

Connection hints — preconnectSitesPassing LCPLCP
0–0 129,717 82% 1.5s
1–1 19,666 78% 1.6s
2+ 31,516 83% 1.6s
Good Needs Improvement Poor Faded rows: under 100 sites

Connection hints — preconnect 0. p75 1. p99 6. At the low end (0–0): LCP 1.5s. At the high end (2+): LCP 1.6s. computed

The State of Web Vitals · Q2 2026 · 189,915 sites · phone field datacorewebvitals.io/state-of-cwv
6

Distribution & median LCP site count and median LCP at each level of connection hints — preload

0ms 750ms 1500ms 2250ms 3000ms
2.5s
0 64262 128523
0–0 1–1 2–2 3–3 4+
Good (≤2.5s) Needs improvement Poor (>4s) Site count
The State of Web Vitals · Q2 2026 · 189,915 sites · phone field datacorewebvitals.io/state-of-cwv
7

Passing LCP by connection hints — preload which level passes the LCP most often

Connection hints — preloadSitesPassing LCPLCP
0–0 128,523 81% 1.6s
1–1 14,628 81% 1.6s
2–2 10,639 86% 1.5s
3–3 4,702 84% 1.6s
4+ 22,407 82% 1.6s
Good Needs Improvement Poor Faded rows: under 100 sites

Connection hints — preload 0. p75 1. p99 18. At the low end (0–0): LCP 1.6s. At the high end (4+): LCP 1.6s. computed

The State of Web Vitals · Q2 2026 · 189,915 sites · phone field datacorewebvitals.io/state-of-cwv
8

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

Hints move network work earlier. dns-prefetch resolves a domain name before it is needed. preconnect goes further and completes the TCP and TLS setup too. preload fetches one specific file that the scanner would otherwise find late, like a background hero image or a critical font.

Hints are easy to overuse. Every preconnect spends sockets and CPU on a connection you might never use. A preload that is not critical steals bandwidth from something that is. A handful of deliberate hints beats a wall of them.

How does this affect the Core Web Vitals?

Connection hints correlate with the LCP. Dns-prefetch hints separates passing sites from failing sites the most. With few dns-prefetch hints, 85% of sites pass the LCP. With many, 74% do. The decline is gradual. There is no point where sites suddenly start failing.

Related signals Cacheable responses → Last-Modified present → Requests via CDN → HTML size (kB) → Chrome field data from 189,915 sites, representing millions of real page loads · How we measured