The three metrics are Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS). Together, they cover loading, responsiveness, and visual stability. The useful question is not “Do we pass?” It is “Which part of the experience is getting in the user’s way?”
01The thresholds at a glance
To be considered good for most users, aim for the 75th percentile of page visits to meet these targets:
How soon the largest meaningful content in the viewport is rendered.
How quickly the page gives visual feedback after an interaction.
How stable visible content remains while the page is in use.
The 75th percentile matters. Core Web Vitals are designed around field data, so a fast experience for the best-connected visitors is not enough. The goal is for at least three out of four page visits to meet the target, assessed separately for relevant device groups.
02LCP: when the page becomes useful
Largest Contentful Paint measures when the largest meaningful text block or image in the initial viewport is rendered. In practice, it often reflects the moment a visitor can see the article headline, product image, hero section, or main visual they came for.
A slow LCP usually has one or more of these causes:
- A slow initial document response.
- A redirect chain or a delayed resource request.
- A large, unoptimised, or incorrectly sized hero image.
- Render-blocking CSS or fonts.
- JavaScript that delays rendering.
- An LCP resource that is loaded too late or lazily loaded by mistake.
Identify the actual LCP element, then trace its path through the network waterfall and rendering timeline. Avoid generic image optimisation until you know whether the image is the problem.
03INP: whether the page stays ready to react
Interaction to Next Paint measures the time from a user interaction to the next visual feedback. It considers click, tap, and keyboard interactions across a visit, not just the first input.
That is important because a page can look complete while still feeling broken. A filter might hesitate, a menu may open late, or a checkout button may appear to ignore a tap because the main thread is busy parsing code, running a long task, or calculating layout.
Common INP fixes include:
- Reducing JavaScript shipped on the initial route.
- Splitting large tasks into smaller pieces.
- Simplifying costly event handlers.
- Avoiding repeated layout reads and writes that force recalculation.
- Reducing hydration work.
- Auditing analytics, chat, experimentation, and other third-party scripts.
INP is a field metric. Lab tools can reveal long tasks and high total blocking time, which are useful diagnostic signals, but real-user data is needed to know whether interactions are consistently responsive in production.
04CLS: whether the page stays put
Cumulative Layout Shift measures unexpected movement of visible content. It captures the frustrating moment when a button jumps just as someone tries to tap it, or when an article moves while the visitor is reading.
The usual causes are not mysterious:
- Images, video, or iframes without reserved dimensions.
- Asynchronous ads, embeds, consent notices, or chat elements inserted above content.
- Font changes that reflow text after the page begins rendering.
- Dynamic UI that pushes existing content down without a user action.
The fix is usually to reserve space, use stable dimensions and aspect ratios, handle font loading intentionally, and avoid inserting content above the visitor’s current position. A page does not need to be static. It needs to be predictable.
05What Core Web Vitals do not tell you
They do not measure every valuable part of an experience. They do not replace accessibility, content quality, security, reliability, conversion measurement, or task-specific performance. A search results page may need a custom metric for when results appear. An ecommerce site may need to measure add-to-cart feedback. A SaaS product may need to track the time until a dashboard is usable.
They also do not give you an automatic SEO result. Strong Core Web Vitals support a better page experience, but they do not compensate for poor relevance, weak content, or an unhelpful page.
06A sensible CWV workflow
- Use Search Console, CrUX, or RUM to find the poor metric and affected page group.
- Segment by device and template before assuming the issue is site-wide.
- Reproduce an example in Lighthouse, DevTools, or WebPageTest.
- Fix the root cause, not every audit suggestion.
- Retest in the lab, then validate the field-data trend.
- Add a budget or release check once the result is stable.
Core Web Vitals are most useful when they lead to a better experience, not a better-looking dashboard.
07FAQ
What are the three Core Web Vitals?
LCP measures loading of the main visible content, INP measures responsiveness to interactions, and CLS measures unexpected layout movement.
What is a good Core Web Vitals score?
At the 75th percentile of page visits, aim for LCP of 2.5 seconds or less, INP of 200 milliseconds or less, and CLS of 0.1 or less.
Did INP replace FID?
Yes. INP replaced First Input Delay as the Core Web Vital for responsiveness because it reflects interactions across a visit rather than only the first one.
Why is my Lighthouse result good but my field data poor?
Your test conditions may be more optimistic than the real world. Field data includes slower devices, weaker networks, different cache states, and interaction patterns that a page-load test may not reproduce.
Which Core Web Vital should we fix first?
Start with the poor metric affecting the most important pages or journeys. Within that group, fix the issue with the clearest user impact and the most evidence of a root cause.