What are Core Web Vitals in plain words?

They're three Google metrics that put numbers on a site's convenience for a person: did the main thing appear quickly (LCP), does the site respond quickly to clicks (INP), and does the content stay still during loading (CLS). A site passes the check when all three are in the green zone for 75% of real visitors.

How do I measure site speed - PageSpeed Insights, GTmetrix or Lighthouse?

Start with PageSpeed Insights (pagespeed.web.dev): it's free, shows all three Core Web Vitals metrics and specific recommendations. GTmetrix is for deeper diagnostics through a waterfall diagram, and Lighthouse is handy for checking pages closed to public access. Always check the mobile version - Google ranks by it.

What are LCP, INP and CLS, and which values count as good?

LCP (when the main element appears) is good under 2.5 seconds. INP (how fast the site responds to actions) is good under 200 milliseconds. CLS (layout jumps) is good under 0.1. INP replaced the old FID metric in March 2024 and measures the response to all interactions in a session, not just the first.

How does site speed affect conversions and SEO?

Conversions, directly: according to Google, when load time grows from 1 to 3 seconds the probability of a bounce rises by 32%, and from 1 to 5 seconds by 90%. So part of the paid ad traffic leaves without waiting for the page. SEO, through the page experience signal: all else being equal, a faster, more stable site ranks higher.

How do I speed up a WordPress, Shopify or WooCommerce site?

Start with compressing images and converting them to WebP - the biggest win on any platform. For WordPress, add caching and remove excess plugins. On Shopify, compress images before uploading and trim heavy apps. For WooCommerce, the main thing is server-side caching and database query optimisation.

Development

Core Web Vitals and Site Speed: How to Make Your Website Faster and Lift Your SEO

Why the same page brings sales to a competitor and sits idle on your site, even though the text and prices are similar. We break down the metrics Google sees and what a business owner should do about them.

A customer arrives from an ad, sees a blank screen, waits one second, then two. On the third they close the tab and go to a competitor whose site opened instantly. You already paid for that click. The money is gone, there’s no lead, and in the report it just looks like “a high cost per conversion.”

A slow site rarely kills a business with one blow. It grinds it down drop by drop: it raises the cost of a lead in advertising, drags the site down in search, and bleeds away traffic you’ve already paid for. The worst part is you don’t see it, because you visit your own site from a cached browser where everything flies.

Core Web Vitals is a set of metrics Google uses to measure how convenient a site is for a real person. In 2021 they became an official ranking signal, and in 2024 one of the three metrics changed. Let’s break down what these numbers are, how to check them in five minutes, and what to do if they’re red.

Why site speed is money, not a technical detail

Owners often treat speed as a question for the tech people: “let the developers sort it out.” The problem is that the bill for a slow site doesn’t go to the developers – it goes to the marketing budget.

The logic is simple. Every second of load time is a slice of people who didn’t wait. A Google study of mobile pages showed that when load time grows from 1 to 3 seconds, the probability of a bounce rises by 32%. From 1 to 5 seconds, it’s already 90%. This isn’t a linear story; the collapse happens fast. Google publishes the details of its methodology openly on web.dev, the resource of the team that develops these metrics.

Now let’s translate that into your money. You pay, say, $1 per click in Google Ads. A person arrives, but the site takes 4 seconds to open, and every third one leaves without waiting. In effect you paid for the click, but the site itself bled away half of those clicks before the visitor even saw the offer. The ads work, the landing page doesn’t. In the report it looks like “expensive traffic,” though the root is technical.

With organic search it’s the same story, just slower. Google directly factors page experience into ranking. All else being equal, between two sites the faster and more stable one ranks higher. On competitive queries this is often the deciding difference between the first and second page of results.

There’s one more layer owners almost never connect to speed: the cost per click in advertising. In Google Ads, your bid is influenced by Quality Score, and one of its components is the landing page experience. In plain words: if your page is slow and inconvenient, Google considers it a worse landing page and makes you pay more for the same click than a competitor with a fast site. The result is a double hit: a slow site both scares people away and makes every click more expensive. Speed up the page and the bid can drop with no changes to the campaigns themselves.

What Core Web Vitals are in plain words

Core Web Vitals are three metrics Google uses to put numbers on what a person feels intuitively: fast or slow, smooth or jerky, convenient or annoying.

Before 2021, Google assessed technical speed in the abstract. Then the company took a step toward the human: instead of “how much do the files weigh,” it started asking “what does the user feel while the page loads.” Hence three simple questions:

Did the main thing appear quickly? – that’s what LCP measures.
Does the site respond quickly when I tap? – that’s what INP measures.
Does everything stay still under my finger while it loads? – that’s what CLS measures.

That’s it. Everything else (shrinking images, caching, deferring scripts) is the toolkit you use to fix those three questions. The metrics are the goal, the tools are the means.

Three metrics: LCP, INP and CLS

Each of the three metrics has clear thresholds from Google: “good,” “needs improvement” and “poor.” A site is considered to pass Core Web Vitals when all three metrics are in the green zone for 75% of real visitors.

Metric	What It Measures	Good	Poor
LCP - Largest Contentful Paint	How quickly the largest visible element (banner, heading, image) appears on the page	up to 2.5 s	over 4 s
INP - Interaction to Next Paint	How quickly the website responds to clicks, taps, and text input	up to 200 ms	over 500 ms
CLS - Cumulative Layout Shift	How much the page content shifts during loading	up to 0.1	over 0.25

LCP - "when I finally saw the main thing"

This is the moment the largest meaningful element appears on screen: a big banner, the heading, the main product photo. Not when the page started loading, but when the person saw the thing they came for. The most common culprit behind poor LCP is a heavy uncompressed image in the header or a slow server response.

INP - "I tapped, and it's thinking"

This is the new metric. In March 2024 Google replaced the old FID (First Input Delay) with it. INP is stricter: it measures the response delay not on the first click, but on all interactions in a session – buttons, filters, add to cart. If you tap “Buy” and the button responds half a second later, that’s poor INP. The usual culprits are heavy scripts: chats, analytics, third-party widgets that block the main thread.

CLS - "I meant to tap one thing and tapped another"

A familiar situation: you’re reading, and suddenly a banner loads at the top, the text jumps down, and you accidentally tap the wrong thing. That’s a layout shift. The most common cause is images and ads without set dimensions, plus fonts that load late and “shove” the content around.

Why these three metrics matter even more for mobile-first markets than they seem. The bulk of traffic comes from mobile devices, and often not from a new flagship on fast Wi-Fi, but from a mid-range phone on a mobile connection that comes and goes. What opens instantly on your office MacBook loads very differently on a customer’s phone on the move. That’s exactly why Google looks at field data from real devices rather than a lab test. And it’s why an owner judging their site from their own fast laptop systematically overestimates its speed.

How to measure site speed: PageSpeed, Lighthouse, GTmetrix

Good news: to see your numbers, you don’t need a developer. There are three free tools, and each is useful for its own thing.

PageSpeed Insights - where to start

Go to pagespeed.web.dev, paste the page address, click “Analyze.” In a minute you get a score from 0 to 100 separately for mobile and desktop, all three Core Web Vitals metrics, and a list of specific recommendations with an estimate of how many seconds each one saves. This is the main tool, because it shows both field and lab data at the same time (more on the difference below). Be sure to check the mobile version: Google ranks sites by it, and most traffic comes from phones anyway.

Lighthouse - the same thing, in the browser

Lighthouse is built into Chrome (a tab in the developer tools). It’s the same engine as under PageSpeed, but runs locally. Handy when you need to quickly check a staging page or one that’s closed to public access.

GTmetrix - when you need to dig deeper

GTmetrix gives a detailed waterfall diagram: you can see exactly which file, script or request is slowing the load and how much each one weighs. It’s a diagnostic tool for when PageSpeed said “slow” and you want to understand what specifically is to blame.

Lab data vs field data: why the numbers differ

This is where owners often get confused, so it’s worth stopping on. PageSpeed shows two types of data, and they can diverge.

Lab data is a one-off test in sterile conditions: an emulation of a single device, a stable connection. Convenient for diagnostics, because it gives the same result on repeat. But it’s not what your real customers see.

Field data is statistics from real Chrome users over the last 28 days, from real phones and real networks. Google collects it in the CrUX report (Chrome User Experience Report) and it’s exactly this that it factors into ranking. If a page is new or has little traffic, there may simply be no field data – then you rely on the lab data.

A practical rule: for SEO decisions, look at field data; to find a specific technical problem, look at lab data. When they diverge sharply, it’s a signal there’s a “bottleneck” somewhere that only shows up on real, slow devices.

«People most often come to us not with “I have a bad LCP.” They come with “why are the ads set up well but the leads are few.” Half of those cases end up not on the advertising side, but on the site side: it simply doesn’t manage to show the offer while the person is still interested. Speed isn’t about technology, it’s about how many paid clicks reach the form».

– Volodymyr Kashalaba, CEO, Guild of Marketing

Common causes of a slow site

Over years of working with client sites, we see that 80% of speed problems come down to a short list of causes. Here they are, from the most common to the rarest.

Heavy, uncompressed images. The most widespread cause. A product photo at 4000×3000 pixels weighing 5 MB, displayed on the site at 400 pixels. The browser still has to download the whole file. The fix is compression and modern formats (WebP, AVIF).
A pile of third-party scripts. A chat, two analytics counters, the Facebook pixel, a reviews widget, a contact-grabbing popup. Each pulls its own code, and together they block the site’s responsiveness (they ruin INP).
Cheap or overloaded hosting. The server “thinks” for a long time before serving the first response. This hits LCP before the images even start loading.
An overloaded theme or builder. Off-the-shelf themes with a hundred features, of which you use five, drag the whole code along “just in case.”
No caching. The site rebuilds the page from scratch each time instead of serving a ready saved version.
Fonts and banners without dimensions. The main cause of layout jumps (poor CLS).

How to speed up WordPress, Shopify and WooCommerce

The approach depends on the platform. One thing unites them: start with the images, because that’s the biggest and fastest win.

WordPress

The most room for optimisation, because the platform is flexible. The basic set: a caching plugin (WP Rocket, LiteSpeed Cache), automatic compression and conversion of images to WebP, lazy loading for images below the fold, minification of CSS and JS. Separately, a review of plugins: you can often disable half of them with no loss. If the theme is heavy, sometimes it’s faster to switch to a light one than to treat the old one.

Tilda

Here you have less control, because the platform is hosted, but the basics are doable: compress images before uploading them, trim heavy apps that each inject their own scripts, remove unnecessary animations and video backgrounds, and choose a light, well-built theme. The ceiling on a hosted platform is set by the platform itself – you can’t jump above it, and that’s worth factoring in when choosing.

OpenCart

Online stores are the hardest, because there are many pages and each one pulls the database. Here server-side caching works, along with database query optimisation, a CDN to serve images from the server closest to the customer, and clearing out unnecessary plugins. On a large catalogue, the difference comes precisely from work with the database, not the images.

Don’t want to dive into the technical details? We run a site speed check with specific recommendations: what exactly is slowing it down, how many seconds you can shave off, and what to do first. Not a generic PageSpeed report, but a prioritised list of actions tailored to your site.

Where to start: the order of fixes

When PageSpeed hands you a list of twenty recommendations, an owner’s hands drop. Good news: you don’t have to do everything at once. The Pareto principle works here too – 20% of the actions give 80% of the result. Here’s the order in which we tackle optimisation, from the biggest win to the smallest.

Images. Almost always number one. Compress, convert to WebP, set dimensions, lazy-load the ones below the fold. This hits two metrics at once: LCP (because the main photo loads faster) and CLS (because images with set dimensions don't jump). Often this alone moves a page from the red zone into the yellow.
Caching. The cheapest win by effort-to-result ratio. The site stops rebuilding the page on every visit. Directly improves LCP.
Third-party scripts. An audit of what's actually needed. A chat nobody messages, two analytics counters instead of one, a widget forgotten from last year. Remove the excess and defer the loading of the rest. Cures INP.
Fixing layout shifts. Set dimensions for all images, banners and ad blocks. Preload fonts so the text doesn't "jump." Targeted fix for CLS.
Hosting and server. If after all the above the server still "thinks" for a long time before the first response, it's time to look at hosting or server-side caching. This is deeper work, but without it you can't break through the ceiling.

The logic is simple: first take what gives the biggest win for the least effort, and only then dig deeper. Tuning the server while uncompressed 5 MB images hang on the site is a waste of money.

Mistakes in DIY optimisation

Owners and marketers often take on speeding up the site themselves – install a plugin, off they go. Sometimes it helps, sometimes it makes things worse. Here are the rakes they step on most often.

Chasing the number 100 instead of real customers. Someone sets the goal “I want 100 points” and spends weeks polishing the last five points that no visitor will feel. Meanwhile the mobile version, where most of the traffic sits, stays slow. PageSpeed is a tool, not an end in itself. What matters isn’t the score, but what a real person feels on their phone.

Several caching plugins at once. A classic WordPress mistake: installing two or three plugins “just in case.” They conflict with each other, and the site starts serving broken pages or breaks entirely. There should be one cache, configured for your setup.

Aggressive minification that breaks the layout. You turn on “combine and minify all CSS and JS” and the site falls apart: buttons don’t click, the slider doesn’t spin, the form doesn’t submit. This is the sneakiest one, because an owner might not notice the breakage on the home page while it sits on the checkout page, exactly where the money is lost. After any optimisation, click the entire customer path to the end.

Optimised the home page, forgot the rest. PageSpeed checks one page by default. The owner gets the home page into the green zone and considers the job done. Meanwhile the product pages, cart and blog stay slow – exactly where conversion happens. You have to check the key page types, not just one.

Compressed photos “to the max” and killed the quality. The overcorrection in the other direction: images compressed so hard the product looks blurry. In a store this directly hits sales – people don’t buy what they can’t see well. Speed matters, but not at the cost of the very thing the person came for.

When people come to us with “the site is slow, we’ve already tried everything,” in nine cases out of ten we open the home page and see the same thing – a four-to-five-megabyte hero banner that loads first and holds the whole page hostage. Before that the owner had installed three caching plugins, minified scripts, dug into the server, while the biggest brake hung uncompressed in plain sight. So we start not with hosting settings, but with the heaviest thing visible on screen right away. The most expensive speed problems are almost always the simplest – nobody just looked at them first.

Who speed is critical for, and who can relax

Honestly: 90 points in PageSpeed isn’t a matter of life and death for every business. Let’s set priorities by type.

Critical – online stores. Here speed converts directly into money. Every second of delay on a product page or in the cart means abandoned carts and lost revenue. On a large catalogue, the difference between a fast and a slow site is easily measured in percentage points of revenue. If you run e-commerce, speed is always a priority.

Critical – sites running paid traffic. If you drive ads to a landing page or site, a slow landing page bleeds budget before the person even sees the offer. Here speed pays off the fastest: optimise the page and the same ad budget brings more leads without a cent extra.

Important, but without fanaticism – content and informational sites. A blog, media outlet, portal. Speed affects SEO and read-through, but there’s no instant “cart abandoned” here. It’s enough to keep the metrics in the green zone, without chasing every last point.

You can relax – a simple brochure site for an offline business. If the site is essentially an electronic sign that a dozen people a day visit to check the address and phone, chasing a perfect INP makes no sense. It’s enough for the page to open normally on a phone and not make people wait. The money for deep optimisation is better spent elsewhere here.

The point isn’t for everyone to have a perfect result. The point is for speed to match the site’s role in the business. Where the site earns – invest. Where it just exists – don’t overpay.

Коли вистачить доробки, а коли потрібен редизайн

This is the main decision an owner with a slow site faces. Spend money on patching, or on a new site? Here’s the benchmark.

A tune-up is enough if the site is modern, the structure is sound, the design suits you, and the problems are pointed: heavy images, excess scripts, no cache. This is fixed with a site tune-up in a few days without rebuilding everything. The best result-to-investment ratio.

You need a rebuild if the site was made 5-7 years ago on an outdated, heavy theme, the code is tangled, every fix breaks something else, and speed is just one of many problems alongside an inconvenient mobile version and weak conversion. In that case, optimising the old one is like repairing an engine that’s already past its service life. Often a site redesign from scratch on a light foundation works out both cheaper and more reliable than endless patching.

If you’re planning a new site, build speed into the technical spec from the start. It’s cheaper to build it fast from the beginning than to optimise it after launch. The same goes for SEO – it’s better to build it in at the site creation stage than to glue it on later.

And one last thing: speed isn’t a one-time event. A site grows new pages, promotions, banners, scripts – and in half a year it’s slow again. So it’s wise to keep Core Web Vitals monitoring within website support, to catch a problem before Google and the visitor do. And if you want the whole picture of “where I’m bleeding money,” speed comes in as a separate block in a marketing audit.

Site is slow and you don't know where to start?

We'll check your site's speed, find exactly what's slowing it down, and tell you honestly: is a tune-up enough, or is a rebuild smarter. No upselling what you don't need.

Volodymyr Kashalaba CEO Guild Of Marketing

Marketing specialist with over 11 years of experience, with a background as a sales department director. Founder and CEO of the Guild of Marketing.

FAQ

What are Core Web Vitals in plain words?

They're three Google metrics that put numbers on a site's convenience for a person: did the main thing appear quickly (LCP), does the site respond quickly to clicks (INP), and does the content stay still during loading (CLS). A site passes the check when all three are in the green zone for 75% of real visitors.
How do I measure site speed - PageSpeed Insights, GTmetrix or Lighthouse?

Start with PageSpeed Insights (pagespeed.web.dev): it's free, shows all three Core Web Vitals metrics and specific recommendations. GTmetrix is for deeper diagnostics through a waterfall diagram, and Lighthouse is handy for checking pages closed to public access. Always check the mobile version - Google ranks by it.
What are LCP, INP and CLS, and which values count as good?

LCP (when the main element appears) is good under 2.5 seconds. INP (how fast the site responds to actions) is good under 200 milliseconds. CLS (layout jumps) is good under 0.1. INP replaced the old FID metric in March 2024 and measures the response to all interactions in a session, not just the first.
How does site speed affect conversions and SEO?

Conversions, directly: according to Google, when load time grows from 1 to 3 seconds the probability of a bounce rises by 32%, and from 1 to 5 seconds by 90%. So part of the paid ad traffic leaves without waiting for the page. SEO, through the page experience signal: all else being equal, a faster, more stable site ranks higher.
How do I speed up a WordPress, Shopify or WooCommerce site?

Start with compressing images and converting them to WebP - the biggest win on any platform. For WordPress, add caching and remove excess plugins. On Shopify, compress images before uploading and trim heavy apps. For WooCommerce, the main thing is server-side caching and database query optimisation.

Monamour

11:46 17 Jan 25

Допомогли створити точку на карті та налаштували контекст. Роботою задоволена, тепер набиваю базу постійних клієнтів.

Барановський Олег

11:28 28 Nov 24

Порекомендували друзі. Почали співпрацю по налаштуванню пошукової реклами в Google, працюємо вже 4-й місяць тому вирішив написати відгук. Подобається підхід в налаштуванні реклами та постійний зворотній звʼязок особливо із питань отриманих лідів, надання звітів щомісяця та обшрунтованих відповідей на мої запити щому сьогодні стільки витратили, на скільки поповнити реклами і тому подібне. Також в нашій ніші недобросовісні конкуренти і нам почали склікувати рекламу, це питання ми також вирішили з акаунт-менеджером і Google нам повернув кошти та далі із цим боремося та моніторимо. Взагалі не знав що таке може бути..
Тому працюємо далі і плануємо покращити сайт та збільшувати бюджет на рекламу.
Рекомендую!

Old School

12:19 12 Mar 24

Я дуже сумнівався чи наймати агенцію чи ні, але вирішив спробувати, бо все ж яка не яка гарантія є. Працюємо вже десь пів року. За цей час мені поправили сайт, результати є, навіть трохи перевиконали

Alex87 M

16:18 27 Feb 24

Супер,тільки позитивні враження, команда професіоналів свого діла, вирішила усі мої питання стосовно гугл акаунтів реклами, рекомендую, буду звертатись ще! Дякую.

Андрій Шевчук

09:52 06 Feb 24

Нам треба було збільшити кількість продаж на сайті і ми вирішили звернутися в цю компанію. За період співпраці ми ще запустили товарну рекламу в гугл та соцмережах. Ми задоволені, що нам не просто налаштовують рекламу, а роблять цілу стратегію залучення покупців, як догнати їх, якщо вони нічого не купили, дають рекомендації і завжди готові відповідати на наші питання.

Віталій Романчук

18:28 03 Feb 24

Займаємось ремонтом квартир, шукав спосіб отримати клієнтів в періоди простою. Родичі порекомендували Гільдію Маркетингу. Не було сайту, тому запустили рекламу в фейсбуці та інстаграмі. Люди пишуть я їм телефоную. Робота є. Замовив ще сайт, очікую що роботи буде ще більше, дякую.

11:32 31 Jan 24

Ми маємо власного маркетолога в команді, але вирішили протестити підрядників. Замовили рекламу в гуглі. Спочатку нам провели презентацію, все показали і порахували що ми отримаємо і скільки треба бюджету. Домовилися про перший пробний місяць. Результат нас влаштовує, заявок стало дійсно більше. Будемо працювати далі, сподіваємося далі отримати ще кращі результати.

Core Web Vitals and Site Speed: How to Make Your Website Faster and Lift Your SEO

Contents

Why site speed is money, not a technical detail

What Core Web Vitals are in plain words

Three metrics: LCP, INP and CLS

LCP - "when I finally saw the main thing"

INP - "I tapped, and it's thinking"

CLS - "I meant to tap one thing and tapped another"

How to measure site speed: PageSpeed, Lighthouse, GTmetrix

PageSpeed Insights - where to start

Lighthouse - the same thing, in the browser

GTmetrix - when you need to dig deeper

Lab data vs field data: why the numbers differ

Common causes of a slow site

How to speed up WordPress, Shopify and WooCommerce

WordPress

Tilda

OpenCart

Where to start: the order of fixes

Mistakes in DIY optimisation

Who speed is critical for, and who can relax

Коли вистачить доробки, а коли потрібен редизайн

FAQ

What are Core Web Vitals in plain words?

How do I measure site speed - PageSpeed Insights, GTmetrix or Lighthouse?

What are LCP, INP and CLS, and which values count as good?

How does site speed affect conversions and SEO?

How do I speed up a WordPress, Shopify or WooCommerce site?