Core Web Vitals Explained

The three measures and their pass marks

Each Core Web Vital answers one plain question about a page, and each has a "good" threshold from Google.

These thresholds are judged at the 75th percentile of page loads, split by mobile and desktop. A page passes when it meets all three. In practice, that means three out of four visits should be at least as good as the target, not the average one.

INP replaced a measure called First Input Delay in 2024. If an older guide mentions FID, it describes the earlier standard.

Field data and lab data are different things

Owners often compare one tool with another and get confused. The tools measure different things.

web.dev says only field measurement can capture the full picture, and that lab testing is not a substitute for it. Lighthouse cannot measure INP because it does not interact with the page, so it uses a stand-in called Total Blocking Time. A page with little traffic may show lab results only. Judge a new site by the lab, then check the field data again once visitors have built it up.

How to read a report, with an example

This example is illustrative. A café website shows these mobile field results in PageSpeed Insights.

The page fails overall, because two measures fail. The decision is straightforward. Leave INP alone. Look at the largest image or text block at the top of the page for LCP, and look for images, embeds or banners with no reserved space for CLS. One oversized photo can cause both problems, since it is slow to arrive and shoves text down when it does. Fixing that one image can move two numbers.

What to fix first, by measure

For LCP. web.dev splits loading into four parts: server response, delay before the main image starts loading, the time to download it, and the delay before it is drawn. It suggests roughly 40% of time on the server response, under 10% on the delay before loading starts, about 40% on download and under 10% on rendering, as guidelines, not rules. Its main fixes are:

• Never lazy-load the main image at the top of the page
• Mark the expected main image as high priority, and only one or two images, because more reduces the effect
• Preload that image if it is only referenced from a style or script file
• Cut redirects and serve cached pages from a content delivery network
• Shrink the image file

For CLS. web.dev's fixes are to set width and height or an aspect-ratio on images and video, reserve space for ads and embeds, avoid inserting banners above existing content, choose font settings that limit shifts and animate with transform instead of layout properties.

For INP. web.dev points to long tasks on the main thread. Reduce script loading and evaluation, keep event handlers short, split heavy work into smaller tasks, and keep the page structure small. Start with field data to find which interaction is slow.

In our judgement, the first and cheapest wins on a small business site are usually images and third-party widgets. Compress and resize images, set their dimensions, and remove chat or booking widgets you no longer use.

Do Core Web Vitals affect rankings?

Yes, as one signal among many. Google says Core Web Vitals are the one part of page experience that its ranking systems use directly, and that good scores do not guarantee top rankings. It also says Search will show the most relevant content even if the page experience is sub-par, and that chasing a perfect score only for SEO may not be worthwhile.

Use this rule. If your report shows "poor" on important pages, fix it, because those are the pages where real visitors have the worst experience. If it shows "good", stop. Spend the effort on pages, reviews and local signals. A passing score will not rescue a page that does not answer the search.

Who should fix what

Our technical SEO checklist lists where speed fits among the other checks, and the explainer on what technical SEO is places it in the bigger picture. If the report is poor and you cannot trace the cause, a short review is the sensible next step. See our web design work or our SEO service.

Next step

Open PageSpeed Insights, test your home page and your main service page on mobile, and note whether the field data passes. If it does not, fix the largest image first.