Website Speed Test: PageSpeed + Core Web Vitals

Core Web Vitals, PageSpeed and what actually moves rankings

Performance is a confirmed Google ranking signal — but only at the threshold level. Going from 60 to 90 helps. Going from 90 to 99 rarely moves the needle. The interesting work happens at the bad-to-good transition.

The three Core Web Vitals — Google's official thresholds

Metric	Good	Needs improvement	Poor
LCP (Largest Contentful Paint)	≤ 2.5 s	2.5 - 4.0 s	> 4.0 s
INP (Interaction to Next Paint)	≤ 200 ms	200 - 500 ms	> 500 ms
CLS (Cumulative Layout Shift)	≤ 0.1	0.1 - 0.25	> 0.25

Since March 2024 INP replaced FID as the interactivity vital. This tool still reports TBT (Total Blocking Time) as a lab proxy — useful, but field INP data from Google's CrUX dataset is authoritative.

Mobile vs. desktop — which one matters more

Google uses mobile-first indexing since 2021. Mobile scores rank pages, desktop scores are a UX courtesy. If your mobile score is 40 and desktop is 95, treat the desktop number as decoration and fix mobile.

The GEO angle — performance and AI citations

Generative engines (Perplexity, ChatGPT Search) have tight render budgets when fetching candidate sources. A slow page may be skipped entirely from the citation pool, even if the content is perfect. The 4-second mark seems to be the soft cutoff in 2026 — pages slower than that drop out of cited sources noticeably.

The fastest wins per metric

LCP > 2.5 s → compress hero image (WebP/AVIF), preload it, host on a CDN.
CLS > 0.1 → explicit width/height on every image and ad slot, no late-loading webfonts.
TBT / INP high → defer non-critical JS, drop third-party tracking pixels, lazy-load below-the-fold widgets.

Frequently asked

Three user-experience metrics Google adopted as ranking signals in 2021. LCP (Largest Contentful Paint) measures when the main content becomes visible, target under 2.5 seconds. INP (Interaction to Next Paint), which replaced FID in March 2024, measures how quickly the page responds to clicks and taps, target under 200 ms. CLS (Cumulative Layout Shift) measures how much the page „jumps around" as it loads, target under 0.1. All three are measured on the 75th percentile of real user visits over 28 days.

INP (Interaction to Next Paint) is the latency between a user interaction (click, tap, key press) and the next visible browser update — measured for the worst-case interaction over the entire visit. FID (First Input Delay) only measured the first interaction. INP is stricter because it catches slow reactions late in the session, when React re-renders or heavy JavaScript stalls become noticeable. Google switched from FID to INP as an official Core Web Vital on 12 March 2024.

Google's official thresholds: „good" under 2.5 seconds, „needs improvement" 2.5-4.0 seconds, „poor" over 4.0 seconds. In practice, sites that beat the 2.5 s mark on mobile — where 4G latency and mid-range CPUs punish slow render paths — win a measurable ranking bump in competitive markets. On desktop most sites clear 2.5 s without effort; the interesting benchmark is mobile 75th-percentile field data.

Lab data (Lighthouse, PageSpeed Insights „Lab" tab) is a single simulated run on a controlled device — good for finding regressions, bad for predicting real-user experience. Field data (the „CrUX" tab, from real Chrome users) aggregates the last 28 days of real visits from real devices, real networks, real distances to your server. Google uses field data for ranking. If your lab score is 95 but CrUX is red, the site is fine on your dev machine and slow for users on 4G in rural areas.

Yes, but as a soft signal — Google has been explicit that Core Web Vitals are a „tiebreaker" between pages of roughly equal content quality. A perfect CWV score on a weak article will not out-rank a strong article with mediocre CWV. But in competitive niches where ten similarly-strong articles fight for position 1, the site that hits „good" on all three vitals has a measurable edge. On mobile the effect is stronger than on desktop.

CrUX (Chrome User Experience Report) is Google's public dataset of real-user performance metrics from Chrome users who opted in to sync. It powers the field-data tab in PageSpeed Insights and the Core Web Vitals report in Google Search Console. Data is aggregated to the origin level (or per-URL for high-traffic pages), updated monthly, and covers the 75th percentile of visits over 28 days. Small sites without enough traffic have no CrUX data and get lab-data-only reports.

Both, and they compound. TTFB (Time to First Byte) measures how quickly your server responds — under 800 ms is Google's „good" threshold. LCP is what the user actually perceives. A slow TTFB (server-side render, database query, cold cache) directly delays LCP because nothing can render until the first byte arrives. Optimising TTFB via caching, CDN and slim server-side rendering is often the single biggest LCP win.

Rankmio's speed test runs a Puppeteer-driven headless Chrome measurement from a Frankfurt server — good for spotting server-side regressions and comparing before/after changes. PageSpeed Insights uses Lighthouse on Google's infrastructure and adds CrUX field data. Both are legitimate; Rankmio is faster to iterate against (no rate limits), PageSpeed Insights is closer to what Google Search Console uses for the CWV report. Use both.

Website Speed Test

Full SEO audit?

Core Web Vitals, PageSpeed and what actually moves rankings

The three Core Web Vitals — Google's official thresholds

Mobile vs. desktop — which one matters more

The GEO angle — performance and AI citations

The fastest wins per metric

Frequently asked

More free Rankmio tools

robots.txt Test

Sitemap Tester

Website Crawler

robots.txt Check

Sitemap Check

Schema.org Check

Domain Check

Word Counter

What is my IP?

Password Generator

Ready to optimize your website?