Website Performance: LCP, INP and CLS in Plain English

Website performance affects whether people can read, compare and act without unnecessary delay. Core Web Vitals describe important parts of that experience. They are useful diagnostic measures, but they do not describe every aspect of usability or accessibility.

Understand the three measures

Largest Contentful Paint, or LCP, concerns when the main content appears. Interaction to Next Paint, or INP, concerns responsiveness after interaction. Cumulative Layout Shift, or CLS, concerns unexpected movement.

Google's current good thresholds are LCP at or below 2.5 seconds, INP at or below 200 milliseconds and CLS at or below 0.1, assessed at the 75th percentile. Interpret mobile and desktop evidence appropriately rather than treating one test as every visitor's experience.

Separate diagnosis from real-user evidence

A laboratory test uses controlled conditions and helps locate problems. Field data reflects actual visits over a reporting period. The results can differ because devices, connections, page states and audiences differ.

Record the method, date and tested page. Compare equivalent evidence after a change. Avoid declaring a site fixed because one homepage test produced a higher score.

Investigate the task and the component

For slow main content, inspect server response, critical resources and media. For poor responsiveness, investigate expensive scripts and interactions. For movement, check reserved image space, fonts, banners and late-loading content.

An illustrative test should include a heavy service page and a form or booking route. A quick text article does not represent the whole site. Keep useful content available while reducing unnecessary weight.

Make a targeted repair

Choose a supported cause, make the smallest useful change and repeat the affected tests. Check that consent, forms and integrations still work. Loading a widget later is only an improvement if the eventual interaction remains dependable.

Keep an accessible static alternative where essential media may fail. Avoid hiding information on mobile simply to improve a score.

Connect performance with outcomes carefully

Review successful tasks and relevant enquiries alongside performance trends. Visitor mix, availability and other site changes can affect conversion. A faster page does not guarantee a specific revenue uplift.

Use the technical audit to prioritise work across templates. Explore website design support for an assessment connected to actual customer journeys.

Investigate the slow task rather than the score alone

Choose an affected template and repeat the same practical action: opening the page, interacting with a menu or submitting a form. Use appropriate lab tools to investigate components and field evidence to understand real-user experience where enough data exists. Record device, connection and test conditions.

Look for the responsible element or interaction before applying a generic optimisation bundle. Large images, third-party scripts, rendering decisions and unstable components can require different repairs. Removing a necessary feature without checking the user task may improve a score while harming the site.

For an illustrative layout-shift issue, identify which element moves and why. Reserve suitable space or repair the component, then repeat the affected task. A single fast test run does not establish the distribution of real-user experiences.

Keep a before-and-after technical record and check that forms, media and navigation still work. When discussing business impact, separate a verified performance improvement from an observed change in enquiries. Pricing, demand and campaign mix may also have changed during the period.

Reference

Google's Web Vitals documentation defines the metrics, thresholds and measurement approach.