Google’s Core Web Vitals are three measures of loading, interaction and visual stability. For real-user field data, the “good” targets are LCP at or below 2.5 seconds, INP at or below 200 milliseconds and CLS at or below 0.1 at the 75th percentile. Check PageSpeed Insights for URL or site data, then use a lab test to investigate specific problems.

Updated 5 November 2026. This guide now explains how Google evaluates real-user data and how to separate a field-data result from a lab test.

If you run a small business, a speed score can be hard to interpret. A high number does not tell you which part of a page is slow, and a single test does not describe every visitor’s experience. This guide explains the three Core Web Vitals, where their targets come from and how to decide what to check next.

What are Core Web Vitals?

Core Web Vitals are Google’s measures for three parts of page experience: loading, interaction and visual stability. They use real-world user experience data when available. They are not a complete measure of whether a site is useful, accessible or easy to use.

Google and web.dev define three metrics:

LCP: when the main content appears

Largest Contentful Paint (LCP) measures how long it takes for the largest content element in the visible page area to render. That element can be a large text block, image, or video poster or frame. LCP often reflects when a visitor can see the main content, though it does not measure every part of loading.

  • Good: 2.5 seconds or less.
  • Needs improvement: more than 2.5 seconds and up to 4 seconds.
  • Poor: more than 4 seconds.

INP: how the page responds to interactions

Interaction to Next Paint (INP) measures the delay between a user interaction, such as a tap, click or key press, and the next visual update. It considers interactions during a visit, so one unusually slow button test is not the same as a site-wide field result.

  • Good: 200 milliseconds or less.
  • Needs improvement: more than 200 milliseconds and up to 500 milliseconds.
  • Poor: more than 500 milliseconds.

CLS: whether the layout moves unexpectedly

Cumulative Layout Shift (CLS) measures unexpected movement of visible page content. For example, a late-loading image can push a paragraph down while someone is reading. CLS is a score, not a duration.

  • Good: 0.1 or less.
  • Needs improvement: more than 0.1 and up to 0.25.
  • Poor: more than 0.25.

These ranges classify the metric values; Google assesses real-user field data at the 75th percentile. The web.dev thresholds guide explains the methodology.

What does the 75th percentile mean?

A percentile describes a group of measurements. If a page’s LCP is 2.5 seconds at the 75th percentile, at least three quarters of the measured visits had an LCP at or below that value. The remaining visits may be slower. The target is not a promise that every individual visit will meet the threshold.

PageSpeed Insights uses Chrome User Experience Report (CrUX) field data when it has enough real-user measurements. It reports the previous 28 days of experience and refreshes the data regularly. It may show data for the specific URL, use data for the whole site when that URL has too few measurements, or show no field data if the site also lacks enough measurements. A low-traffic page may therefore have no URL-level field result.

Google Search Console’s Core Web Vitals report groups similar URLs and reports field data when enough measurements are available. Use it to spot patterns across groups of pages, then inspect an example URL in PageSpeed Insights. A group result does not mean every URL in that group has identical performance.

Do Core Web Vitals affect search rankings?

Google says its ranking systems use Core Web Vitals as part of the page-experience signals they seek to reward. Google also says there is no single page-experience signal and that good reports do not guarantee top rankings. Relevant content can still rank when its page experience is imperfect. Treat the metrics as a way to find user-experience problems, not as a ranking guarantee or a substitute for useful content.

For a business site, a performance check can help you identify friction. It cannot predict how many visitors will stay, enquire or become customers. Those outcomes also depend on the visitor’s needs, the offer, the page content and other factors.

Why can a site look fast in one test and slow in another?

Lab tests and field data answer different questions

A Lighthouse lab test runs a page under a simulated device and network setup. It helps a developer reproduce a problem and inspect diagnostic details. PageSpeed Insights can show a Lighthouse lab result alongside field data when field data is available.

Field data comes from real Chrome users and reflects a range of devices, networks and visits during the reporting period. It can differ from a lab test because the conditions differ. Use the lab report to investigate; use available field data to understand how measured visits performed.

Check mobile and desktop data separately

PageSpeed Insights and Search Console can distinguish mobile and desktop experiences. Compare the device category and date range before drawing conclusions. A test on your own phone or office connection is useful for finding a visible issue, but it does not represent every customer’s device or network.

Loading and interaction are different

A page can show its main content quickly and still respond slowly when someone opens a menu or submits a form. LCP and INP measure different parts of the experience, so check all three metrics before choosing a fix.

How can you check your business website?

Start with a page tied to an important customer task or a recurring customer complaint, such as a service or contact page. If Search Console flags a group, check a representative URL in PageSpeed Insights. When similar pages show the same issue, ask the maintainer to check their shared template; when one URL differs, investigate that page content or assets. Then use each tool for the question it answers:

  • PageSpeed Insights: enter a page URL to see available field data and a Lighthouse lab test. If URL field data is missing, check whether PSI shows origin data, which combines measurements across pages on your site, or no field data.
  • Google Search Console: review the Core Web Vitals report for groups of URLs with enough field measurements. Open an example URL in PSI to investigate further.
  • Chrome DevTools Lighthouse: run a lab test while developing or troubleshooting a page. The test can help reproduce issues but does not replace field data.

If neither URL nor origin field data is available, you cannot judge field performance from that report yet. Use Lighthouse to investigate a specific observed issue, and treat its result as a lab diagnostic rather than a real-user assessment.

Record the URL, device category, test date, field-data availability and metric before comparing results. A different date, device or URL can produce a different result.

What should you fix first?

Use the report to locate a bottleneck before changing the site. PageSpeed Insights lists diagnostics for its lab test, while web.dev explains common fixes for each metric. The best next step depends on what the page and report show.

If LCP needs attention

Check which element becomes the largest contentful paint and when the browser discovers and loads it. Depending on the report, useful checks can include:

  • whether the main image is larger than its rendered size;
  • whether the browser discovers the main image promptly;
  • server response time and other delays before rendering;
  • CSS or JavaScript that blocks the main content from appearing.

Image compression or a newer format can help when an oversized image is part of the measured problem. It is not a universal first fix; follow the page’s diagnostics.

If INP needs attention

Inspect the interaction that feels delayed and check whether the page is doing substantial work at the same time. A developer can review long JavaScript tasks, event handlers and third-party scripts. Removing code without checking what the report identifies can break a feature without improving the measured interaction.

If CLS needs attention

Look for content that moves after it first appears. Common checks include reserving space for images or embeds, reviewing late-loading banners and making sure content is not inserted above text the visitor is reading. The cause should be verified on the affected page.

If you do not maintain the site, share the URL, device category, report and a short description of what you observed with the person who does. Ask them to connect the proposed change to a specific diagnostic. After the change, repeat the Lighthouse test with the same URL and device setup to check the lab result. Review PSI or Search Console field data later, when enough post-change visits are available; a lab retest alone does not show real-user experience.

Does the website platform determine performance?

A platform alone does not tell you how fast a site will be. Themes, plugins, scripts, hosting, images and the way a page is built can all affect the result. WordPress can be configured in different ways, just as a custom or static site can include expensive scripts or large assets.

If you are choosing or reviewing a platform, compare the features and maintenance needs your business requires. For performance work, test the actual pages and fix the causes shown by the data. The guide on whether WordPress is right for your small business website covers that broader choice.

What should you take from a Core Web Vitals report?

Use the targets as a consistent way to interpret field data: LCP at or below 2.5 seconds, INP at or below 200 milliseconds and CLS at or below 0.1 at the 75th percentile. Check whether PSI has URL, origin or no field data, and use Lighthouse diagnostics to investigate a specific issue.

A good result does not guarantee a ranking or an enquiry. If a report identifies a problem on a page that matters to your business, you can describe what you found and ask for help prioritising the next technical check.