Core Web Vitals explained without the acronyms

Someone has sent you a report with three coloured bars and a set of initials, and asked for a decision. Here is what the three things measure, what counts as passing, and how much you should actually care.

They measure three different kinds of annoying

Every one of these is a stopwatch pointed at a specific way a website irritates the person using it. That is the whole idea. Strip the names off and they are three complaints a customer would make out loud.

Largest Contentful Paint: is anything there yet

The complaint: I tapped your link and stared at a white screen.

This measures how long it takes for the biggest visible thing on the screen to finish loading, usually a hero image or a headline. Google's threshold for good is 2.5 seconds, measured on real visits from real people, most of them on a phone on mobile data rather than your office wifi.

On local business sites the culprit is nearly always an image. A photo straight off a phone camera is several megabytes and gets shrunk to the size of a business card by the browser after it has already downloaded the whole thing.

Interaction to Next Paint: did that do anything

The complaint: I tapped the menu and nothing happened, so I tapped it again, and then it opened twice.

This measures the delay between someone tapping and the page visibly responding. Good is 200 milliseconds or less. It replaced an older measurement in 2024 that only looked at the very first interaction, which flattered sites that were fine to arrive at and painful to use.

The usual cause is the quantity of third-party scripts loaded into the page: chat widgets, booking embeds, review sliders, two analytics tools, a heatmap recorder someone installed for a trial in 2023 and never removed.

Cumulative Layout Shift: stop moving

The complaint: I went to tap the phone number and an advert loaded and I tapped that instead.

This measures how much the content jumps around while the page is loading. Good is 0.1 or below, on a scale where 0 is a page that never moves. It is the one that most directly costs money on a local site, because the thing people are reaching for is usually the call button, and a call you never received is the one number no report will show you.

How much does this matter for a local business

Less than the person selling you a performance retainer says, and more than nothing.

Google has confirmed these are used as a signal and has never published how much weight they carry. Anyone quoting you a percentage is making it up. In practice, the effect we see is that page speed rarely wins a competitive local search on its own, and regularly loses one, because a site that takes eight seconds on a phone loses the visitor before ranking has a chance to matter. Set it against what does move the local algorithm and it is a long way down the list.

Nobody has ever chosen a plumber because the site loaded quickly. Plenty have given up on one because it did not.

What to actually do about it

1. 01Resize your photos before uploading. Most site images never need to be wider than 2000 pixels, and a photo out of a phone is far larger.
2. 02Remove third-party scripts you are not using. Go through them one at a time and ask what each is for. Two or three will have no answer.
3. 03Give images and embeds a fixed size in the page so nothing reflows around them as they arrive. It belongs with the rest of your on-page work rather than in a separate speed project.
4. 04Check your own site on your own phone, off wifi, with a stopwatch. It is a cruder measure than the report and it tells you more.
5. 05Then stop. Beyond passing, this is not where a local business's next customer comes from.

Speed is one part of the technical layer. The wider version, including the parts that decide whether your pages are found at all, is on our technical SEO page. If your site is fast and you are still invisible, the problem is more likely to be somewhere in how the local pack is won.