Core Web Vitals: What They Are and How to Fix Them
By The seo.bike team, SEO & GEO practitioners · Last reviewed July 8, 2026
Core Web Vitals are three user-experience metrics: LCP measures loading (target under 2.5 seconds), INP measures responsiveness to input (target under 200 milliseconds), and CLS measures visual stability (target under 0.1). They are a real but modest ranking signal, mainly a tiebreaker between otherwise comparable pages.
LCP: what it measures and how to fix it
Largest Contentful Paint marks the moment the biggest visible element in the viewport, usually a hero image, a video poster, or a large block of text, finishes rendering. It is the searcher's felt sense of "the page has loaded." Good is under 2.5 seconds, needs-improvement is 2.5 to 4 seconds, and poor is over 4 seconds, measured at the 75th percentile of real visits.
The common causes are a slow server response, render-blocking CSS and JavaScript, and heavy unoptimized images. The fixes map directly: cut server time with caching and a CDN, defer or inline critical CSS so the browser paints sooner, and compress and correctly size the LCP image, serving modern formats and a proper srcset so mobile does not download a desktop-sized file.
Two high-leverage moves: preload the LCP image so the browser fetches it early instead of discovering it late, and avoid lazy-loading the hero image, which is a frequent self-inflicted LCP wound. Lazy-loading is for below-the-fold images; applying it to the largest above-the-fold element delays the very metric you are trying to improve.
INP: the responsiveness metric
Interaction to Next Paint replaced First Input Delay in 2024 and measures how quickly the page visually responds to a user's interactions, taps, clicks, and key presses, across the whole visit, reporting roughly the worst interaction. Good is under 200 milliseconds, needs-improvement is 200 to 500, and poor is over 500.
A sluggish INP almost always traces to JavaScript hogging the main thread: long tasks that block the browser from responding while they run. The fixes are about breaking up and deferring that work. Split long tasks into smaller chunks that yield to the browser, remove or defer third-party scripts that run on every interaction, and avoid heavy work in event handlers.
Practical wins include debouncing expensive handlers, moving non-urgent computation off the main thread or into idle time, and trimming the amount of JavaScript shipped in the first place. Because INP looks at your worst interaction, one heavy handler on a common button can sink the whole page's score even if everything else is snappy.
CLS: visual stability
Cumulative Layout Shift measures how much visible content jumps around unexpectedly as the page loads. You have felt it: you go to tap a button and an ad loads above it, shoving the button down, and you tap the wrong thing. CLS is scored on the movement, and good is under 0.1, needs-improvement is 0.1 to 0.25, and poor is above 0.25.
The usual culprits are images and videos without declared dimensions, ads and embeds that inject themselves with no reserved space, and web fonts that swap and reflow the text. Each has a clean fix: always set width and height (or a CSS aspect-ratio) on media so the browser reserves the box before the file arrives, reserve fixed space for ad and embed slots, and load fonts so they do not cause a large reflow.
The subtle one is content inserted above existing content after load, such as a cookie banner or a lazy-loaded section pushing everything down. Reserve its space up front or insert it below the fold. CLS is often the cheapest Vital to fix because the causes are structural and the remedies are one-time CSS changes.
Field data vs lab data
There are two ways to see these metrics, and they answer different questions. Field data (also called real-user monitoring, and what Google's Chrome User Experience Report collects) is measured from actual visitors on their real devices and networks. This is what Google uses for the ranking signal, aggregated at the 75th percentile over 28 days.
Lab data is a single simulated load in a controlled environment, like the test a tool runs on demand. It is repeatable and great for debugging because it isolates one page under fixed conditions, but it does not reflect the spread of real devices, and notably it cannot measure INP, which requires real interactions to observe.
The practical consequence: fix using lab data because it is reproducible and diagnostic, but judge success using field data because that is what ranks and what users experience. A page can score beautifully in a lab run and still fail in the field because most of your visitors are on mid-range phones and congested networks the lab never simulated.
Where to measure them
Start with Search Console's Core Web Vitals report, which groups your URLs into good, needs-improvement, and poor buckets using field data across your whole site, so you see which page templates fail at scale rather than one URL at a time. This is the right entry point because problems usually cluster by template.
For a single URL, PageSpeed Insights shows both the field data (if the page has enough traffic) and a fresh lab run with specific opportunities. When a page lacks enough traffic for field data, you fall back to the lab score and the CrUX origin-level data as a proxy.
For debugging, the browser's built-in performance and Lighthouse tooling lets you trace exactly which element is the LCP, which task blocks INP, and which node shifts for CLS. Fix in the browser tools, confirm the fix landed for real users weeks later in Search Console, since field data lags because it averages 28 days of visits.
How much do they actually move rankings?
Here is the honest part. Core Web Vitals are a confirmed ranking signal, part of the page-experience signals, but they are a lightweight one. Google has said they act mainly as a tiebreaker: when two pages are otherwise comparably relevant, the better experience can win. They will not lift irrelevant or thin content over strong, relevant content.
So do not expect a poor-to-good CLS fix to rocket a page up the results. The larger returns from improving these metrics are usually indirect: faster, more stable pages reduce bounces, lift conversions, and improve how many pages get crawled, all of which help the business even where the direct ranking nudge is small.
The sensible posture is to get out of the "poor" bucket, because a genuinely bad experience can hold a page back and it frustrates real users, but not to chase a perfect score at the expense of content and relevance, which move rankings far more. Treat Vitals as hygiene and a conversion lever, not as a growth strategy on their own.
Key terms
Related guides
- How Search Engines Work: Crawling to Ranking
- How to Implement Hreflang Without Errors
- How to Use Canonical Tags Correctly
- Technical SEO: The Crawl, Index & Speed Foundations
Frequently asked questions
They are LCP (Largest Contentful Paint) for loading, with good under 2.5 seconds; INP (Interaction to Next Paint) for responsiveness, with good under 200 milliseconds; and CLS (Cumulative Layout Shift) for visual stability, with good under 0.1. Each is measured at the 75th percentile of real visits, so most of your visitors must hit the target for the page to pass.
Interaction to Next Paint (INP) replaced First Input Delay (FID) as the responsiveness metric in March 2024. INP is stricter and more representative: rather than measuring only the delay of the first interaction, it evaluates the responsiveness of interactions across the whole visit and reports roughly the worst one, which is why heavy JavaScript event handlers now show up as a problem.
Field data comes from real visitors on their own devices and networks and is what Google uses for ranking, aggregated over 28 days at the 75th percentile. Lab data is a single simulated test in a controlled environment, repeatable and good for debugging but not reflective of real-world spread, and it cannot measure INP. Fix with lab data, judge success with field data.
They are a real but minor ranking signal that mostly acts as a tiebreaker between otherwise comparable pages; they will not push irrelevant or thin content above strong, relevant content. The bigger payoff is indirect, through lower bounce rates, better conversions, and improved crawling. Aim to escape the poor bucket, but do not chase a perfect score over relevance.