Google considers your site fast enough when three measurements score well for real visitors. The largest content appears within 2.5 seconds (LCP), the page responds within 200 milliseconds (INP) and the layout shifts by at most 0.1 (CLS), for three out of four visits, separately for phone and computer [1].
Getting a website built with AI? Ask for a measurement instead of a promise. One red score in PageSpeed Insights says less than you think; below you can read why, and which fixes deliver the most.
The three measurements in plain language
| Metric | What it measures | Good | Poor |
|---|---|---|---|
| LCP, Largest Contentful Paint | When the largest visible element, usually your main image or heading, is on screen | ≤ 2.5 s | > 4 s |
| INP, Interaction to Next Paint | How quickly the page visibly responds to a click, tap or key press | ≤ 200 ms | > 500 ms |
| CLS, Cumulative Layout Shift | How much the layout shifts unexpectedly while loading | ≤ 0.1 | > 0.25 |
In everyday terms: LCP is the moment your visitor thinks "the page is there". INP is the feeling that a button is lagging. CLS is the button that jumps away just as you want to click it.
A page only passes if all three measurements are good [5]. On 12 March 2024, INP replaced the old metric First Input Delay (FID) [6]; articles that still mention FID are outdated.
Start with these five fixes
You often come across these causes on SME sites, and they are relatively quick to fix:
| Fix | Improves | Effort |
|---|---|---|
Shrink the main image, use WebP or AVIF, don't lazy-load it, add fetchpriority="high" [7], [8] | LCP | Low |
| Width and height on every image and video [9] | CLS | Low |
| Clean up third-party scripts: remove what you don't use, load the rest as late as possible [10] | INP, LCP | Low to medium |
| Reserve space for the cookie banner, embeds and promo bar [9] | CLS | Low |
Self-host fonts, limit them to what you use, set font-display [9] | LCP, CLS | Medium |
In our experience, that is where most of the gain lies for SME sites. Only then come the bigger jobs: breaking up long JavaScript tasks, since a task of more than 50 milliseconds counts as long [11], or replacing a heavy page builder. Why a static site suffers less from this is covered in WordPress or a static website.
If your own code is optimised but your site stays slow, web.dev says the cause is probably third-party scripts [10]. Some of your advertising and analytics pixels can run through your own server. Whether that helps, and what it costs, Cédric explains in server-side tracking explained.
Real visitors or lab test: which figures count?
PageSpeed Insights shows two blocks. At the top is field data from the Chrome User Experience Report (CrUX): the experience of real Chrome users over the past 28 days [5]. The Core Web Vitals report in Search Console relies on it too [12].
Below is the lab test (Lighthouse): a single simulated page load, for mobile with a mid-range device on a mobile network [5]. It produces the score from 0 to 100, a weighted mix of five lab metrics [13]. Useful for finding causes, but Google itself does not expect a perfect 100 [13].
Many SME sites see "no data" at the top. That is not an error: a page only enters CrUX when it is "sufficiently popular", Google does not disclose the threshold and you cannot register your site yourself [14]. Chrome on iOS doesn't count either [14].
In that case, use the lab test as a guide, not as a report. And test it yourself on an ordinary phone, not on office wifi.
What speed does for your position in Google
"Core Web Vitals are used by our ranking systems", Google writes. But good results don't guarantee a top position, and according to Google, chasing a perfect score purely for SEO may not be the best use of your time [15]. Google always shows the most relevant content, even if the page experience is weaker [15].
Our reading: speed is a precondition, not a trump card. It matters more to your visitors: at Vodafone, a 31% better LCP produced 8% more sales [16]. That is a big player and a case from 2021, but the direction holds for an SME too.
And the web as a whole? In 2025, 48% of mobile websites had good Core Web Vitals. For most, LCP is the stumbling block: 62% achieve a good LCP on mobile, compared with 77% for INP and 81% for CLS [17].
How to stay fast after launch
A new site is fast on the day it is delivered. The question is whether it stays that way, after ten new photos, a chat widget and an extra pixel.
- Every month: test your five most important pages in PageSpeed Insights on mobile, and check the Core Web Vitals report in Search Console.
- After an improvement: be patient. The field data covers 28 days [5], and if you click "Validate fix" in Search Console, Google monitors the group for 28 days [12].
- With every new script or widget: test the page before and after.
What speed costs at build time is covered in how much does a website cost in 2026? Want to know how your current site scores on a phone, and what is causing it? The free website scan measures loading speed, Core Web Vitals and mobile display.
Frequently asked questions
What are good Core Web Vitals?
LCP within 2.5 seconds, INP of 200 milliseconds or less and CLS of 0.1 or less, for 75% of page visits, separately for mobile and desktop. A page only passes if all three are good.
What is the difference between the PageSpeed Insights score and the Core Web Vitals?
The score from 0 to 100 comes from a lab test with a simulated device and network. The assessment at the top comes from real visits by Chrome users over 28 days, if your site has enough traffic. Search Console relies on those real figures.
Why does PageSpeed Insights show no data for my site?
Because your page or site has too few visits in the Chrome User Experience Report. Google does not disclose the threshold and you cannot register your site yourself. That leaves the lab test, and you had best also test on an ordinary phone yourself.
Will I rank higher with a faster website?
A little at most. Google uses Core Web Vitals in its ranking systems, but says that good scores don't guarantee a top position and that the most relevant content comes first. Speed mainly helps your visitors, and with them your enquiries.
What is INP?
Interaction to Next Paint: how quickly your page visibly responds after a click, tap or key press. INP replaced First Input Delay (FID) as a Core Web Vital on 12 March 2024. Articles that still talk about FID are outdated.
Sources
- Web Vitals — web.dev (Google), 31 Oct 2024, accessed 27 Sep 2026
- Largest Contentful Paint (LCP) — web.dev (Google), 4 Sep 2025, accessed 27 Sep 2026
- Interaction to Next Paint (INP) — web.dev (Google), 2 Sep 2025, accessed 27 Sep 2026
- Cumulative Layout Shift (CLS) — web.dev (Google), 12 Apr 2023, accessed 27 Sep 2026
- About PageSpeed Insights — Google for Developers, 21 Oct 2024, accessed 27 Sep 2026
- Interaction to Next Paint becomes a Core Web Vital on March 12 — web.dev (Rick Viscomi), 12 Mar 2024, accessed 27 Sep 2026
- Optimize Largest Contentful Paint — web.dev (Google), 31 Mar 2025, accessed 27 Sep 2026
- Choose the right image format — web.dev (Google), 13 Aug 2024, accessed 27 Sep 2026
- Optimize Cumulative Layout Shift — web.dev (Google), 7 Feb 2025, accessed 27 Sep 2026
- Load Third-Party JavaScript — web.dev (Google), 19 Feb 2024, accessed 27 Sep 2026
- Optimize long tasks — web.dev (Google), 19 Dec 2024, accessed 27 Sep 2026
- Core Web Vitals report — Search Console Help, accessed 27 Sep 2026
- Lighthouse performance scoring — Chrome for Developers, accessed 27 Sep 2026
- CrUX methodology — Chrome for Developers, 20 Jun 2024, accessed 27 Sep 2026
- Understanding page experience in Google Search results — Google Search Central, 22 Sep 2026, accessed 27 Sep 2026
- Vodafone: A 31% improvement in LCP increased sales by 8% — web.dev (Google), 17 Mar 2021, accessed 27 Sep 2026
- Web Almanac 2025: Performance — HTTP Archive, 5 May 2026, accessed 27 Sep 2026