Every now and then a client gets an email from an SEO agency: “Your Core Web Vitals are poor, please get in touch.” Most website owners delete it on the spot. They shouldn’t. You can understand these three metrics in ten minutes, and it’s worth it, because Google really does use them for ranking.
Core Web Vitals are measurements taken from real visits to your website. Not from a lab, not from someone’s computer at an SEO agency. From the phones and connections of your actual visitors. Google collects them through the Chrome browser and aggregates them. That makes them more honest than any one-off test run on a fast connection.
There are three of them.
LCP: how long until the content shows up
Largest Contentful Paint is the moment a visitor sees the largest element on the page, usually the header image or the main headline. Target: under 2.5 seconds. Over 4 seconds is poor.
What usually drags this score down? A huge hero image weighing four megabytes. A server that makes you wait half a second for the first response. Scripts that block the page from rendering.
The quickest fix I know: resize images to the size they are actually displayed at and switch to modern formats. That alone can cut LCP by a second.
INP: does the page respond when I click
Interaction to Next Paint measures the time between a click and the page’s response. You hit “send” and watch whether the button reacts or the page just freezes and stares back blankly. Target: under 200 milliseconds. People don’t notice delays shorter than that, so it’s the threshold at which a page “works instantly”.
The usual culprits are ad scripts, heavy animations and too many plugins. On sites with three cookie plugins, two chat widgets and a map in the background, INP can come out four times too high. Each of those scripts is fast on its own. Together they fight over the same resources.
CLS: does the page jump around under my finger
Cumulative Layout Shift measures how much page elements move around while it loads. The classic scenario: you’re reading an article on your phone, the page pulls in a banner, the text jumps down and you tap the wrong thing. Target: under 0.1.
The troublemakers: images without set dimensions, banners pushed into the content after loading, fonts swapped halfway through rendering. The cure is simple and tedious: every image needs a declared width and height, and space for ads and banners has to be reserved up front.
How to check it
Go to PageSpeed Insights, paste in your website’s address and read the results. It’s free and needs no login. Focus on the real-user data section, not the colourful scores from the simulation: the first is the truth, the second is only a model.
The second option: Google Search Console has a dedicated Core Web Vitals report, if your site is verified there. It shows how many visits met the thresholds and how many didn’t.
What this has to do with how we build websites
Static websites, like the ones Webomat produces, have a naturally easy time with these metrics. No heavy content management system, no ten plugins, images sized correctly from the start. It isn’t magic; there are simply fewer things that can break.
How to read a PageSpeed Insights report
The report has two parts, and people mix them up all the time. The top one, “Discover what your real users are experiencing”, is data from actual visits, collected by browsers. The bottom one, “Diagnose performance issues”, is a simulation on a single reference phone.
The top part is the truth about your website. The bottom part is an estimate, albeit a very useful one: it shows what to fix. For a new site with little traffic, the top section is often empty because there isn’t enough data. In that case, look at the bottom one and treat it as a rough guide.
The second thing hardly anyone mentions: the scores in the circles (the green, orange and red ones from 0 to 100) are aggregates. For ranking, Google uses the percentage of visits that meet the thresholds, not the circle score. You can have a score of 70 with healthy vitals, or a score of 90 with one serious problem. Look at the percentages, not the circles.
The most common culprits and how to fix them
Across the few dozen audits we’ve done, the same list keeps coming up:
A huge header image. A photo straight from the camera, 6,000 pixels wide, four megabytes. Fix: resize it to screen width and save it in a modern format. Result: LCP drops by as much as two seconds. It’s the cheapest fix in the whole audit.
Dozens of third-party scripts. A chat widget, two analytics tools, a map, a cookie banner, an ad pixel. Each one is fast on its own; together they block the main thread. Fix: audit what you actually use and get rid of the rest. Usually two or three survive.
Images without set dimensions. The browser doesn’t know how much space to reserve, so the content jumps down. Fix: add a width and height to every image. An hour’s work, and CLS drops close to zero.
Fonts loaded too late or twice. Text flickers or looks different for a second. Fix: load the right font files early and use one set across the whole site.
A sluggish server. The first response arrives after 800 milliseconds, before anything starts to render. Fix: faster hosting or server-level caching. No trick in the code will help here.
What it costs to have someone fix it
Optimising an existing website is a developer’s job. Typically:
- review and report: two to four hours,
- image optimisation and small fixes: two to four hours,
- dealing with scripts and fonts: two to six hours,
- something bigger, like changing how the page loads: up to fifteen hours or so.
At $75 to $150 an hour, the simple stuff comes to $300 to $1,000. The heavier work, $1,000 to $2,500. Before you pay, check whether the problem is the server itself: switching to $100-a-year hosting can be more effective than a week of development.
When to stop fixing
You know the point at which renovating an old flat stops being worth it? Websites are the same. If the site’s engine is ten years old, the plugins are held together with tape and every fix breaks two other things, optimisation just prolongs the agony. Replacing it costs less, and the result lasts for years.
That’s why, before any fixing, it’s worth knowing what can be saved. A free screenshot and audit shows your starting point in a minute: speed, mobile version and number of pages. That’s where the decision starts: renovate or replace.
Got a website on an old system and not sure where to start? The Webomat audit measures all of this for free and shows you the result in a minute.