Key Takeaways
- Core Web Vitals are three Google measurements of how a page feels to use — how fast it appears, how quickly it responds to taps, and how much it jumps around while loading.
- 53% of mobile visitors abandon a page that takes longer than 3 seconds to load (Google, "The Need for Mobile Speed," 2016).
- A 0.1-second improvement in load speed lifted retail conversion rates by 8.4% across 37 major brands (Deloitte/Google, "Milliseconds Make Millions," 2020).
- You don't need to read code to check your own score — one free Google tool gives you a plain pass/fail per page.
- Slow, unstable pages also get crawled and cited less by AI tools like ChatGPT and Google's AI Overviews, not just ranked lower by traditional search.
If your website feels sluggish, or a page jumps as it loads, that's not a cosmetic issue — it's a Google ranking factor, and it's actively costing you customers. Core Web Vitals are three specific, measurable signals Google uses to judge whether a page is fast and stable enough to be a good experience: how quickly the main content appears, how quickly the page responds when someone taps a button, and how much the layout shifts around while it loads. You don't need to understand a line of code to grasp why they matter — you just need to have waited for a slow website to load on your phone.
This guide translates the three metrics into plain English, shows you what a slow site actually costs in real customers, and gives you a free way to check your own score in under two minutes.
What are Core Web Vitals, in plain English?
Core Web Vitals are Google's three headline measurements of real-world page experience: LCP (how long the main content takes to appear), INP (how quickly the page responds when someone interacts with it), and CLS (how much the layout shifts unexpectedly while loading). Google has used them as a confirmed ranking signal since May 2021, and they're measured from real visits to your site, not a lab test.
Think of them as three separate questions a mystery shopper would ask about your shop:
- Is the shop actually open when I arrive, or am I stood at a locked door? That's LCP.
- When I ask a member of staff a question, do they answer straight away, or do I have to wait? That's INP.
- Is the till fixed to the floor, or does it keep sliding away from me as I try to pay? That's CLS.
None of these are abstract engineering concerns. They're the exact moments a real customer notices, gets frustrated, and leaves — they just have technical names when Google measures them at scale.
Why does a slow website cost you customers?
A slow website loses you customers before they've read a single word of your copy, because most visitors decide within a few seconds whether to stay or leave. 53% of mobile site visits are abandoned if the page takes longer than 3 seconds to load (Google, "The Need for Mobile Speed," 2016). That single statistic, now a decade old, is why speed became a boardroom issue and not just a developer one.
The relationship isn't a cliff edge at 3 seconds — it's a steady decline that starts from the first second. Portent's analysis of over 100 million page views across 20 B2B and B2C websites found e-commerce conversion rates fall from 3.05% at a 1-second load time to 0.67% at 4 seconds (Portent, "Site Speed is (Still) Impacting Your Conversion Rate," 2022) — a drop of more than three-quarters. Every additional second is measurably fewer enquiries, fewer bookings, fewer sales, on the exact same traffic you're already paying to attract.
The upside runs the same direction. When Deloitte and Google studied 37 leading retail, travel and luxury brands across more than 30 million user sessions, they found that a 0.1-second improvement in load speed increased retail conversion rates by 8.4% and lifted average order value by 9.2% (Deloitte/Google, "Milliseconds Make Millions," 2020). A tenth of a second is barely perceptible to a human being — but it moved revenue at scale, because it moved the outcome for thousands of individual visits, each one either continuing through to a purchase or bouncing off.
What do LCP, INP and CLS actually measure?
Each of the three Core Web Vitals measures a different moment of frustration, and Google has set a specific pass/fail threshold for each one based on what real users tolerate.
| Metric | What it measures | "Good" threshold |
|---|---|---|
| LCP — Largest Contentful Paint | How long until the biggest visible element (usually a hero image or headline) has rendered | 2.5 seconds or less |
| INP — Interaction to Next Paint | How long the page takes to visibly respond after someone taps a button, link, or form field | 200 milliseconds or less |
| CLS — Cumulative Layout Shift | How much visible content jumps around unexpectedly as the page finishes loading | A shift score of 0.1 or less |
(Source: web.dev, Google, "Web Vitals," current thresholds, updated October 2024.)
LCP is the one most people mean when they say a site "feels slow" — it's the wait before there's anything real to look at. INP is newer; it replaced an older metric called FID (First Input Delay) as an official Core Web Vital in March 2024, because FID only measured the delay before a tap registered, not whether the screen actually updated in response. INP measures the full round trip: tap, then visible reaction. CLS is the one behind the most infuriating mobile experience of all — tapping a button, only for an advert or image to load in above it and shift the whole page down, so you tap the wrong thing instead.
Google doesn't average these across your whole site. A page only counts as passing if 75% of real visits clear the threshold for all three metrics simultaneously. One weak metric drags the whole page down, even if the other two are excellent.
How fast should my website really be, and how does mine compare?
Most websites currently fail this bar. Google's own Chrome User Experience Report — which measures real visits to real sites, not lab tests — found that only 43% of websites pass all three Core Web Vitals together on mobile (Google Chrome User Experience Report (CrUX), 2024). Individually, 63% pass LCP, 74% pass INP, and 79% pass CLS — but because a page needs all three at once, the combined pass rate drops well below any single metric.
That gap between the individual scores and the combined score is exactly why so many business owners are confused when a developer tells them their site is "mostly fine." Mostly fine on each metric individually still fails the bar Google actually uses.
Checking your own site takes about two minutes and needs no login or code. Go to PageSpeed Insights, paste in your homepage URL, and press enter. It returns a score out of 100 for mobile and desktop, plus a plain-language pass/fail for each of the three Vitals, based on real visitor data where your site has enough traffic to measure it. If you manage Google Search Console, its "Core Web Vitals" report shows the same data grouped by URL, so you can see exactly which pages are dragging the average down.
What actually makes a website slow?
In most small-business sites we've audited, the same handful of causes show up repeatedly, and none of them are exotic:
- Oversized images. A photo taken straight from a phone camera, uploaded without resizing, can be 5–10MB — dozens of times larger than a page actually needs. This is the single most common cause of a poor LCP score.
- Template bloat. Off-the-shelf website builders and page-builder plugins load code for every feature the template could use, not just the ones your page actually uses. That unused code still has to download and run before the page settles.
- Autoplaying video and stacked pop-ups. Both are common CLS culprits — they load in after the visible content and shove everything else out of the way.
- Third-party embeds. Chat widgets, review badges, and ad scripts are usually the biggest contributor to a poor INP score, because they hijack the browser's attention while the user is trying to interact with the page.
- Shared or budget hosting. If your server is slow to respond in the first place, no amount of front-end tidying will fix LCP — the clock starts before your page even begins rendering.
None of these are permanent characteristics of "a website." They're specific, fixable choices — often made by whoever built the site originally, under time pressure, without measuring the outcome.
What can I fix myself, and what needs a developer?
A genuinely useful split for a non-technical owner is: content-level fixes you can often do yourself, and structural fixes that need someone who builds websites.
You can usually handle:
- Compressing images before uploading them (free tools like Squoosh do this in the browser, no installation needed)
- Removing autoplaying video and unnecessary pop-ups
- Cutting plugins or embeds you added once and never look at
- Checking your PageSpeed Insights score monthly, so a regression gets caught before it costs three months of traffic
This usually needs a developer:
- Rebuilding on a lighter, custom-coded foundation instead of a heavy template
- Fixing server response time or migrating to better hosting
- Reserving space for images and ads before they load, to eliminate layout shift at the source
- Restructuring how and when JavaScript loads, to fix a poor INP score sitewide rather than page by page
If your site consistently fails Core Web Vitals across multiple pages, that's usually a sign the underlying build is the problem, not any single image or plugin. That's the point at which patching individual pages stops being worth the time, and a rebuild on Core Web Vitals-optimised foundations from day one — which is what our bespoke web design work is built around — pays back faster than continuing to chase symptoms.
How does site speed affect being found by Google and AI tools?
Site speed doesn't just affect whether a human visitor stays — it affects whether Google and AI search tools can crawl, index, and cite your page in the first place. Core Web Vitals have been a confirmed Google ranking factor since 2021, so two pages with equally good content can rank differently purely on load speed and stability.
That effect has extended to AI search tools too. As we've covered in our guide to GEO vs SEO, AI crawlers respect the same technical access signals traditional search engines do — a slow, unstable page gets crawled less often by the bots that decide what ChatGPT, Perplexity, and Google's AI Overviews are allowed to quote. Speed is table stakes for both games, not just the traditional one. Our broader guide to what generative engine optimisation actually is covers the AI citation side in more depth, and our complete AI search playbook walks through the full 90-day rollout for owners who want to tackle both at once.
Speed alone won't get you cited, though — the content still has to answer the question clearly once it loads. We've written a step-by-step guide to structuring your pages so AI tools can quote them directly: How to Write Answer-First Content for AI Search. Fast and clear beats fast alone.
Where to start: Run your homepage and your two most-visited service pages through PageSpeed Insights today. If any of the three metrics fail, start with image compression — it's the highest-impact, lowest-effort fix available, and it takes under an hour.
Frequently Asked Questions
Do Core Web Vitals really affect Google rankings for a small business site?
Yes. Google confirmed Core Web Vitals as a ranking signal in May 2021, and it applies to every indexed website regardless of size. It won't outweigh genuinely better content on a competitor's page, but between two similarly relevant pages, the faster and more stable one has a measurable advantage. For a small business competing against larger sites with bigger content budgets, this is one of the few ranking factors that's entirely within your control.
What's the quickest free way to check my Core Web Vitals score?
Go to PageSpeed Insights, enter your website's URL, and press enter. Within about 30 seconds it returns a score out of 100 for both mobile and desktop, with a plain pass or fail against LCP, INP, and CLS, plus specific suggestions ranked by impact. No account or code access is needed.
Can I improve Core Web Vitals myself, or do I need a developer?
Some fixes are within reach for a non-technical owner — compressing images before upload, removing autoplaying video, and cutting unused plugins or embeds are the three highest-impact changes most people can make alone. Fixes that require touching how the site is built — server response time, how JavaScript loads, or reserving layout space for ads — need a developer. If your site fails Core Web Vitals across most pages rather than just one or two, that's usually a sign the underlying build needs attention, not the content.
How much does a slow website actually cost a small business?
Google's research found 53% of mobile visitors abandon a page that takes longer than 3 seconds to load (Google, 2016), and Portent's analysis of 100 million page views found e-commerce conversion rates fall from 3.05% at 1 second to 0.67% at 4 seconds (Portent, 2022). Applied to your own traffic and average order value, that's a concrete, calculable revenue loss — not a vague technical concern.
Do Core Web Vitals matter for AI search tools like ChatGPT, not just Google?
Yes. AI crawlers use the same technical access signals as traditional search bots, so a slow or unstable page gets crawled less often by the systems that decide what AI Overviews, ChatGPT, and Perplexity are able to quote. Speed won't get you cited on its own — the content still needs to answer the question clearly — but it's a prerequisite either way. See our guide to GEO vs SEO for the full picture of what's shared between the two systems and what's genuinely different.
Sources
| # | Statistic | Source | Year |
|---|---|---|---|
| 1 | 53% of mobile visits abandoned after a 3-second load | Google, "The Need for Mobile Speed" | 2016 |
| 2 | 0.1s speed improvement lifted retail conversions 8.4% | Deloitte/Google, "Milliseconds Make Millions" | 2020 |
| 3 | E-commerce conversion falls from 3.05% (1s) to 0.67% (4s) | Portent | 2022 |
| 4 | Core Web Vitals "Good" thresholds (LCP 2.5s, INP 200ms, CLS 0.1) | web.dev, Google | 2024 |
| 5 | 43% of websites pass all three Core Web Vitals on mobile | Google Chrome User Experience Report (CrUX) | 2024 |
| 6 | INP replaced FID as an official Core Web Vital | web.dev, Google | 2024 |