Core Web Vitals explained: what your host fixes and what it can't
Three metrics decide whether Google calls your pages fast. Your host owns part of one of them. The rest is on your site.
The three metrics, briefly
Largest Contentful Paint (LCP) measures how long it takes for the biggest thing in the viewport, usually a hero image or headline block, to finish rendering. Google wants it under 2.5 seconds.
Interaction to Next Paint (INP) measures how quickly the page responds when someone taps or clicks, across the whole visit rather than just the first interaction. Under 200 milliseconds is the target.
Cumulative Layout Shift (CLS) measures how much the page jumps around while loading. A score under 0.1 counts as good. Ads, late-loading fonts and images without dimensions are the usual culprits.
What your host genuinely controls
Server response time is the honest answer, and it is a component of LCP rather than the whole thing. Time to First Byte (TTFB) is how long the server takes to start sending the page. Slow shared hosting, an overloaded database, or a data centre on the wrong continent all push TTFB up, and every millisecond there delays everything that follows.
Hosting also decides your PHP version and available memory, whether server-level caching exists, whether HTTP/2 or HTTP/3 is offered, and whether a CDN is bundled. Those matter, and a good host makes several of them a non-decision.
If your TTFB is above roughly 600 milliseconds for cached pages, the host is a real part of your problem. Below that, moving hosts will not save you.
What hosting cannot fix
INP is almost entirely a JavaScript problem. It measures how long the browser's main thread is blocked when a user interacts, and that blocking comes from your scripts: tag managers, chat widgets, analytics, heavy page builders, third-party ad code. No host on earth can make a 400 KB blocking script cheap to execute on a mid-range Android phone.
CLS is a markup problem. Images and iframes without width and height, web fonts that swap in a different size, banners injected above content after paint. A faster server delivers the same shifting layout, just sooner.
And most of LCP is usually your own doing too: an unoptimised hero image, a render-blocking stylesheet, a font that must download before the headline can paint.
The order that actually helps
Start with the image in your hero. Serve it in WebP or AVIF, size it for the viewport rather than shipping a 3000-pixel-wide original, and mark it as high priority so the browser fetches it early. This one change moves LCP more than anything else on a typical site.
Then audit third-party scripts. Load each one and ask what it earns. Defer everything that is not needed for first paint, and delete the ones nobody has looked at in a year. This is where INP is won.
Then set explicit width and height on every image and reserve space for anything injected later. That is CLS handled.
Only then look at your host. If TTFB is still poor after caching is in place, the platform is the constraint.
Field data versus lab data
Google ranks on field data, collected from real Chrome users and published in the Chrome UX Report. Tools like PageSpeed Insights show both that field data and a lab test run on a simulated device.
The lab score is a diagnostic, not the thing being measured. A site can score 98 in the lab and still fail Core Web Vitals in the field because real visitors are on slower phones, worse networks, and pages the lab test never loaded.
Field data also updates on a 28-day rolling window, so a fix made today will not show up for weeks. Change one thing at a time or you will not know what worked.
Where a host upgrade is worth it
Move when your cached TTFB stays high, when traffic spikes take the site down, or when you are on a plan whose CPU limits throttle you during normal business hours. Those are platform problems and no amount of image optimisation solves them.
In our hosting tests, Hostinger (9.6/10) and SiteGround (9.3/10) were the two providers whose server response times stayed flat under sustained load, which is the property that matters for LCP.
If your TTFB is already low and your scores are still failing, the fix is on your side of the wire. Moving hosts will cost you a weekend and change nothing.
Frequently asked questions
Do Core Web Vitals actually affect rankings?
Will a CDN fix my Core Web Vitals?
Why does PageSpeed Insights disagree with my own testing?
How long after a fix do the scores update?
Is shared hosting automatically bad for Core Web Vitals?
Related reading
HostingWhat Is TTFB? Time to First Byte Explained for 2026
TTFB measures how long your server takes to start responding, here is what counts as good and how to test yours properly.
HostingHow to Speed Up Your Website in 2026, Ordered by Impact
Hosting quality, caching, image compression, and a CDN, ranked by how much faster each one actually makes your site.
HostingLiteSpeed vs NGINX vs Apache in 2026: Plain-English Guide
The web server your host runs shapes how fast pages leave the building, here is the difference without the jargon.