Blog · Insights

Core Web Vitals: Complete Guide to LCP, INP & CLS

Core web vitals measure loading, interactivity, and visual stability. Learn what LCP, INP, and CLS mean, how they affect rankings, and how to improve them.

If someone has told you your site is slow, but not what to actually fix, this is for you. Core web vitals are Google's way of measuring how your pages feel to a real person: how fast the main content loads, how quickly the page responds when someone taps a button, and whether things jump around while it loads. That is the whole idea, stripped of jargon. In this honest breakdown we explain what LCP, INP and CLS mean in plain words, whether speed actually affects your sales, and what is genuinely worth fixing versus what is not worth touching.

We work with small businesses across the Cyprus market: Paphos hotels, Limassol service firms, local shops with an online catalogue, and the same questions come up every time. Here is the simple way we explain it to clients, with no slides and no scare tactics.

What core web vitals actually measure

Core web vitals are a small set of measurements Google uses to score the experience of loading and using a web page. They are part of what Google calls page experience signals, and they sit alongside other things like whether your site works on mobile and uses HTTPS. There are exactly three of them right now: loading, interactivity, and visual stability.

The important part, and the part most articles skip, is that these scores come from real visitors, not a test you run once. Google collects anonymous performance data from people using the Chrome browser who have opted in. This is called field data, and it is what counts for rankings. A separate kind of test, called lab data, runs in a controlled environment and is useful for debugging. The two often disagree. This is exactly why your developer can run a quick test that looks fine while your actual visitors, on older phones and patchy mobile signal in a Paphos cafe, get a worse experience.

Google sorts each metric into three buckets: Good, Needs improvement, and Poor. It judges this at the 75th percentile, which sounds technical but just means this: take the slowest quarter of your visitors and ignore them, then look at the rest. If the bulk of your visitors get a fast, stable page, you pass.

One useful detail: Google has stated these metrics will not change more than once a year, and any changes are announced clearly. So this is not a moving target you need to chase weekly. According to Google's Chrome User Experience Report data, fewer than half of all websites globally pass all three Core Web Vitals thresholds at once, which gives you a sense of how much room for improvement most sites still have.

If you are wondering how this fits into the bigger picture of why some sites simply do not show up in search, our piece on why company sites are invisible in search covers the full range of factors, of which performance is one part.

LCP, INP and CLS in plain words

Here are the LCP INP CLS metrics, one at a time, in language you can actually use. These are the three numbers behind every "your site is slow" conversation.

LCP: how fast the main thing shows up

Largest Contentful Paint (LCP) measures how long it takes for the biggest visible element to load. On most pages that is a hero image, a banner, or a big block of text. A good LCP is 2.5 seconds or faster. If a visitor stares at a mostly blank screen for four or five seconds, that is a poor LCP, and that is usually what people mean when they say a site "feels slow".

The most common cause of a slow LCP on small business sites is a large, uncompressed hero image. The browser has to download the whole file before it can paint it, and on a mobile connection that gap is painful. Compressing that image and serving it in a modern format like WebP is usually the single fastest win available. If your site is on WordPress, a caching plugin that also handles image delivery through a content delivery network (CDN) handles most of this automatically. Our notes on SEO for small businesses in Cyprus go into which hosting and caching setups tend to work well for sites in this market.

INP: how quickly the page reacts to you

Interaction to Next Paint (INP) measures responsiveness: when someone taps a button, opens a menu, or clicks "add to basket", how long before the page visibly responds? A good INP is 200 milliseconds or less. INP recently replaced an older metric called First Input Delay (FID). The difference is simple: FID only looked at the very first interaction, while INP looks across all interactions during a visit, which is a fairer picture of how the page actually behaves.

CLS: whether things jump around

Cumulative Layout Shift (CLS) measures visual stability. You have probably tried to tap a link, then an image loaded above it, the page shifted, and you tapped the wrong thing. That is layout shift, and it is infuriating. A good CLS is 0.1 or lower. The usual culprits are images without set dimensions, ads or banners that load late, and fonts that swap in after the page appears.

So when you ask what are good core web vitals thresholds, the short answer is this: LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1. Pass all three on most visits and you are in good shape.

How a slow site quietly loses customers

Let us answer the obvious question first: does website speed actually affect sales, or is this just something agencies say to sell work? Honestly, it depends on how bad things are. If your site loads in three seconds and your competitor loads in 2.8, nobody is making a buying decision based on that. Diminishing returns are real, and we will tell you when you have hit them.

But a genuinely slow site loses customers in ways you never see. Someone clicks your link, waits, gets bored, and goes back to Google before your page even finishes loading. You never know they were there. It is like pouring water into a bucket with holes: you can spend more on ads to drive traffic, but a slow site lets a chunk of that traffic leak straight back out.

This matters most on mobile, which is where most local searches happen. A tourist looking up restaurants in Paphos on their phone, on hotel wifi or roaming data, has very little patience. The slower and shakier your page, the higher your bounce rate, the lower your engagement, and the fewer enquiries you get. This is the heart of web performance optimization: a faster, more stable page tends to keep people there long enough to actually do something.

There is also the question of how core web vitals affect Google rankings. They are a real ranking factor, but they are a tie-breaker, not the main event. Google has been clear that relevance and useful content come first. Two pages that are equally useful, one fast and one slow, will see the faster one favoured. A slow page with the best answer will still often outrank a fast page with a weak one. So the honest framing is this: page experience signals help good content perform better; they do not rescue thin content.

For a broader view of how speed fits into local search performance, our local SEO checklist for Cyprus sets out the full range of things worth checking, performance included.

What is worth fixing, and what is not

This is where most money gets wasted. People get a red score, panic, and pay for fixes that move a number nobody cares about. Here is what is worth doing and what is not.

Usually worth fixing:

• Oversized images. The single most common cause of slow loading on small business sites. Compressing and correctly sizing images is boring and effective, and it improves LCP fast.
• No caching. If your site rebuilds every page from scratch for every visitor, it will feel sluggish. A caching setup is a one-time win that helps everyone.
• Layout shift from unsized images and late-loading banners. Cheap to fix, and it removes a genuinely annoying experience.
• Heavy or unnecessary scripts. Tracking tags, chat widgets, and plugins you forgot you installed all slow interactivity and hurt INP.

Usually not worth chasing:

• Squeezing a score from "good" to "slightly better good". Once you pass the threshold, the extra effort rarely pays back.
• A perfect 100 lab score. The lab number is for debugging. Field data is what counts, and a perfect lab score does not guarantee a good field score.
• Chasing a single bad metric on a page nobody visits. Fix the templates that affect many pages and the pages that actually bring traffic and revenue first.

The smart approach to web performance optimization is to fix things at template level. If your product page template is slow, fixing it once improves hundreds of pages. That is far better value than tweaking pages one by one. When you work through LCP INP CLS metrics, do it in the order of impact: biggest visitor numbers and biggest revenue pages first.

For a typical WordPress small business site, the single most impactful fix is usually moving hero images and other large assets into a CDN-served pipeline with automatic format conversion to WebP. This addresses LCP at scale across every page that uses those images, in one change, rather than optimising files one at a time.

If you are planning a bigger refresh, it is worth thinking about performance and structure together rather than bolting speed fixes on afterwards. Our notes on website redesign and SEO cover how to do that without losing what already works.

How to check your own site for free, and what to do next

You do not need to pay anyone to find out where you stand. Here is a short guide to the main free tools you can run yourself today.

• PageSpeed Insights: Paste in your URL. It shows both field data (real visitors, the bit that counts) and lab data (for debugging), plus specific suggestions. Start here.
• Google Search Console: Its Core Web Vitals report groups your URLs into Good, Needs improvement, and Poor based on real-world usage, by URL pattern. This is the best view of how your whole site is doing, not just one page.
• Lighthouse and Chrome DevTools: Built into the Chrome browser. Useful for your developer to diagnose specific problems.
• Third-party trackers: Tools like SpeedVitals and DebugBear track vitals over time (28 days, 6 months, a year) so you can see whether a change actually helped.

Here is how to read the results without panicking. Look at the field data first. If most metrics are green, you are fine; do not let a red lab score talk you into spending money. If you see red in the field data, note which metric and which pages, because that tells you what to fix first. If your lab test looks good but your field scores are poor, that usually means real visitors are on slower devices or connections than your test assumed, which is common and worth taking seriously.

Google Search Console also shows you which URL patterns are worst affected, so you can prioritise by traffic rather than guessing. If you are not yet set up with Search Console, our guide to SEO for small businesses in Cyprus walks through the setup alongside the other basics worth getting in place early.

If the reports leave you unsure what is actually broken and what is not worth touching, that is normal. A focused website visibility audit looks at performance alongside everything else affecting your search results, so you spend on the fixes that matter rather than the ones that just look scary.

Frequently asked questions

Core web vitals are not magic and they are not a scam. They measure a few sensible things about how your site feels to real people: speed, responsiveness, and stability. Get the basics solid, oversized images, caching, layout shifts, heavy scripts, and you have covered most of what matters. Beyond that, diminishing returns kick in, and we would rather tell you to stop than sell you fixes that do nothing.

For most WordPress small business sites, the single most impactful starting point is a CDN-served image pipeline with automatic WebP conversion. It fixes LCP across every page at once, costs very little to set up, and is the kind of boring, effective change that actually moves field data. Everything else can follow once that is in place.

If you want a second pair of eyes on what is broken, what is not worth touching, and what to fix first, that is exactly the kind of plain-language conversation we have with clients across Cyprus. The first consultation is free, with no slides and no jargon walls.

Get a free, honest look at your site