Core Web Vitals explained without jargon

The first is about how quickly the largest element appears — usually a photo or a heading. That is the moment a visitor thinks: it is here. What slows it down are oversized images, fonts arriving late and servers answering slowly.

The second is about shifting. Everyone knows it: you want to tap something and the page moves because a banner or image is still loading. That is preventable by reserving space for imagery and adverts before they arrive.

The third is about response: when you tap something, how quickly does something visible happen. That mainly gets bad from too many scripts doing work in the background at exactly the moment the visitor tries something.

Testing tools run a single measurement on a simulated device. The numbers that count come from measurements on real visitors, on their own devices and connections. A site can score excellently in a test and disappoint in the field, for instance because many visitors use older devices.

So look at field data where it exists, and test yourself on an ordinary phone with a mediocre connection rather than on a fast laptop with fibre.

And weigh the outcome: scoring 100 on a page nobody visits is wasted effort, while half a second gained on your busiest page is immediately noticeable.

Images larger than they are displayed on screen. On the vast majority of sites this is the biggest item, and it is the cheapest win: resize, compress and serve in a modern format.

Third-party scripts: chat widgets, video players, ad pixels, social media integrations. Each is an external party slowing your site and watching along. What you genuinely need is loaded so it does not hold up the page.

Fonts that load late, leaving text invisible at first or shifting afterwards. That is fixable with a few lines, and it is one of the most visible improvements.

And finally the system underneath: pages recalculated on every visit while nothing has changed. Pre-rendering where possible often saves more than all other measures combined.