Is page speed a Google ranking factor?

Yes, through Core Web Vitals and the broader page experience signal, though it acts mainly as a tiebreaker between pages of otherwise similar relevance. Content that answers the query still wins over faster content that doesn't. The larger, more reliable payoff is behavioural: faster pages convert better and lose fewer visitors to abandonment, independent of any ranking effect.

What are the three Core Web Vitals?

LCP (Largest Contentful Paint, loading, good at 2.5s or less), CLS (Cumulative Layout Shift, visual stability, good at 0.1 or less), and INP (Interaction to Next Paint, responsiveness, good at 200ms or less). INP replaced First Input Delay as the third vital in March 2024. All three are assessed on real-user mobile data, not just a lab score.

What is the single most common cause of a poor performance score?

An unoptimised hero image combined with render-blocking CSS or JavaScript ahead of it in the head. The image is usually the LCP element, and it cannot paint until the browser has finished downloading and processing whatever blocks it first, so both the image itself and what delays reaching it need fixing together.

Does compressing my images fix a bad LCP score on its own?

Often only partly. If render-blocking resources in the head are delaying when the browser even starts painting, a smaller image will still arrive late, because the wait happens before the image request is the bottleneck. Fix render-blocking resources and image weight together rather than assuming either one alone will close the gap.

Should I test performance on desktop or mobile?

Mobile. Google's assessment is mobile-first, most visits are mobile, and phones have slower CPUs and less reliable connections, so a page that looks fine on a desktop test can still fail in the field on the device most visitors actually use.

How does the Kalenux audit measure Core Web Vitals?

Through Google's own PageSpeed Insights API, run on mobile, checked against the same good and poor thresholds Google publishes for LCP, CLS and total blocking time (the audit's proxy for responsiveness). Alongside that, the crawler records asset weight, compression, minification and render-blocking resources in the head so a poor vital score can be traced back to a specific cause rather than left as a single number.

Home/SEO Guide/Performance

SEO guide · category 07 · experience

2.5 s.
200 ms.
0.1.

Three numbers decide whether Google calls a page fast. They measure how quickly the main content appears, how fast the page responds, and how still it stays while loading.

Last reviewed · 1 Oct 2026Reference article · 25 sources · 1030 words →Kalenux engineering

Test my speed free →PageSpeed Test

/shop/trail/ · real users · 75th percentile

LCPLargest Contentful Paint2.1s

good ≤ 2.5 spoor > 4.0 s

INPInteraction to Next Paint280ms

good ≤ 200 mspoor > 500 ms

CLSCumulative Layout Shift0.02

good ≤ 0.1poor > 0.25

illustrative

one page load, second by second

Each metric is a moment in the same four seconds.

Watching a single load shows why the three metrics measure different things.

Start

TTFB 0.6s

FCP 1.2s

LCP 2.1s

Shift 2.6s

Tap 3.4s

0 s1 s2 s3 s4 s

StartRequest sent

TTFB 0.6sFirst byte arrives from the server

FCP 1.2sFirst text or image is painted

LCP 2.1sMain image appears: this is LCP

Shift 2.6sA banner pushes text down: counts toward CLS

Tap 3.4sPage takes 280 ms to respond: this is INP

illustrative

two kinds of speed data

Real visits or one simulated run.

PageSpeed Insights shows both. They answer different questions, and they rarely agree.

Field datawhat users felt

p75

Thousands of real Chrome visits over 28 days. The value reported is the 75th percentile: three in four visits were at least this fast.

Lab datawhat to debug

1 run · 1 device · 1 network

A single simulated load on a set device and connection. Repeatable and detailed, so it’s where you find causes, but it isn’t what Google’s ranking uses.

what drags each number down

Different metric, different culprits.

Fixing the wrong one is common: a smaller image won’t help a page that’s slow to respond.

LCPMain content is slow

01Slow server response (high TTFB)
02Render-blocking CSS and scripts
03Large, unoptimised hero image
04Content only built by JavaScript

INPPage is slow to respond

01Long JavaScript tasks on the main thread
02Heavy third-party scripts
03Very large DOM to update
04Work done on every keystroke or scroll

CLSLayout jumps around

01Images and video without width and height
02Ads and embeds that load late
03Banners inserted above content
04Web fonts swapping size on load

where the time goes

Read the waterfall top to bottom.

Each bar is one file. The long red bar is the LCP image arriving late because it waits behind the CSS and a font.

document

styles.css

font.woff2

hero.jpg

app.js

analytics.js

0 s1 s2 s3 s

fix: preload hero.jpg and defer app.js · illustrative

performance myths, checked

Three things people believe about page speed.

MYTH

A Lighthouse score of 100 means I’ll rank higher.

WHAT’S TRUE

Google’s ranking uses real-user Core Web Vitals, not the lab score, and good vitals don’t guarantee top positions. Relevance still comes first.

MYTH

If my content is great, speed doesn’t matter.

WHAT’S TRUE

Core Web Vitals are part of Google’s ranking systems, and slow pages lose visitors before they read anything.

MYTH

My score should be the same every time I test.

WHAT’S TRUE

Lab runs vary with network, server load and third-party scripts. Compare trends and field data, not single runs.

Reference article · 25 sources

Want the full picture, with every claim sourced?

Web performance, in search engine optimisation, is how quickly and smoothly a web page loads and responds for the people using it. Google measures it chiefly through Core Web Vitals, three user-centred metrics, Largest Contentful Paint (LCP), Interaction to Next Paint (INP) and Cumulative Layout Shift (CLS), that cover loading, responsiveness and visual stability. The metrics were introduced in 2020 and are used by Google’s ranking systems as part of page experience.

KALENUX REFERENCEWeb performance and Core Web Vitals

1History
2The three Core Web Vitals
3Supporting metrics
4How performance is measured
5Role in Google Search ranking
6Improving LCP
7Improving INP
8Improving CLS
9Common misconceptions

25 references · cite-readyRead the article →

See your Core Web Vitals across every page.

Free to start. Real-user data where it exists, lab data everywhere else.

Start my free audit →

2.5 s.200 ms.0.1.

Each metric is a moment in the same four seconds.

Real visits or one simulated run.

Different metric, different culprits.

Read the waterfall top to bottom.

Three things people believe about page speed.

A Lighthouse score of 100 means I’ll rank higher.

If my content is great, speed doesn’t matter.

My score should be the same every time I test.

Want the full picture, with every claim sourced?

Speed is sorted: make sure every locale gets it.

See your Core Web Vitals across every page.

2.5 s.
200 ms.
0.1.