Page speed is how quickly a web page loads, becomes usable and responds when someone interacts with it. It sounds like a single number, but it is really a group of measurements that describe what a visitor experiences from the moment they click a link until the page settles down.
For business owners, page speed matters for a simple reason: slow pages frustrate people. For SEO and AI search, it is one of the signals that tells Google your page offers a good experience. This guide explains what page speed means, how it is measured and what to fix first.
What is page speed?
Page speed describes how fast the content on a specific URL loads and reacts to visitors. It differs from "site speed," an average across many pages. A lightweight blog post and an image-heavy product page on the same site can perform very differently.
Today, Google measures the parts of page speed that matter most to users with Core Web Vitals. According to web.dev, the three metrics and their "good" thresholds are:
- Largest Contentful Paint (LCP): how long the main content takes to appear. Good is 2.5 seconds or less.
- Interaction to Next Paint (INP): how quickly the page responds after a click, tap or key press. Good is 200 milliseconds or less.
- Cumulative Layout Shift (CLS): how much the layout jumps around while loading. Good is 0.1 or less.
Google assesses these at the 75th percentile of page loads, separately for mobile and desktop. In plain terms, at least three out of four visits should hit the "good" mark.
Why page speed matters for search and AI answers
Google's page experience documentation confirms that Core Web Vitals are used by its ranking systems. It also makes clear that relevance comes first. Google says it will show the most relevant content even when page experience is sub-par, while good page experience can help when many helpful pages compete for the same query.
So speed rarely makes a weak page rank, but it can separate two strong ones. It also shapes what happens after the click, since visitors staring at a blank screen often leave.
The same logic carries into AI search. Google's guidance on AI features such as AI Overviews says there are no special optimizations required, and lists a great page experience among the standard practices site owners should follow. Fast, stable pages are also easier for crawlers to fetch, which supports crawling and indexing across a large site.
How page speed is measured
Field data
Field data comes from real visitors. The Chrome User Experience Report (CrUX) collects anonymized performance data from Chrome users, and PageSpeed Insights shows it based on the previous 28 days. It reflects what your audience actually experiences on their own devices. Low-traffic pages may fall back to site-wide data.
Lab data
Lab data comes from a simulated test, usually run by Lighthouse on an emulated mid-range phone. It is useful for debugging because you can rerun it after each change. However, Lighthouse cannot measure INP, since no real person is interacting with the page, so it uses Total Blocking Time as a stand-in. Lab scores also vary from run to run.
Where to check
The Core Web Vitals report in Google Search Console groups your URLs by status, which helps you spot templates with problems. PageSpeed Insights is best for checking a single URL and seeing specific recommendations.
Page speed in practice
Picture a hypothetical local services company with a homepage built around a large hero photo, a chat widget and several tracking scripts. In PageSpeed Insights, its mobile field data shows LCP around four seconds and INP that "needs improvement," while desktop looks fine.
Digging in, the team finds the hero image is a very large, lazy-loaded file, so the browser discovers it late. The chat widget and tag scripts load in the page head and delay taps on the menu. Fixing those two issues addresses what real mobile visitors are hitting.
How to improve page speed
Start with field data to find the failing metric. The guides on web.dev for LCP, INP and CLS go deep, but these steps cover most sites.
Speed up the main content (LCP)
- Compress images and serve them at the right size in modern formats like WebP or AVIF.
- Make the main image discoverable in the initial HTML and give it a high fetch priority if needed. Never lazy-load the image that forms the largest element.
- Reduce server response time with caching and a content delivery network (CDN), and cut unnecessary redirects.
- Remove unused CSS and defer non-critical JavaScript so it does not block rendering.
Make interactions feel instant (INP)
- Audit third-party scripts such as chat widgets, heatmaps and ad tags. Remove ones you do not use.
- Break long JavaScript tasks into smaller pieces so the browser can respond to input in between.
- Keep the page's DOM (its HTML structure) reasonably small, since large pages are slower to re-render.
Stop layout jumps (CLS)
- Add width and height attributes to images and videos so the browser reserves space.
- Reserve space for ads, embeds and banners, or overlay them instead of pushing content down.
- Choose fallback fonts that closely match your web fonts to reduce text reflow.
Common page speed mistakes
- Chasing a 100 Lighthouse score. The lab score is a diagnostic tool. Field data and Core Web Vitals reflect real users.
- Only testing on desktop. Google indexes the mobile version of your pages, and mobile devices on slower networks usually struggle most. See mobile-first indexing.
- Lazy-loading everything. Lazy-loading below-the-fold images is helpful. Lazy-loading the hero image delays LCP.
- Assuming speed alone will lift rankings. Speed supports good content, it does not replace it.
How Butter approaches page speed
At Butter, we treat page speed as part of a page's overall health, not a vanity score. On our GEO+SEO plan, technical SEO fixes are included, and slow templates, render-blocking scripts and layout shifts are the kind of issues we flag and work through alongside broken links and crawl problems. We look at real-user data in Search Console first, then use lab tools to find the cause. The Butter Analytics dashboard connects to Google Analytics and Search Console, so speed changes can be viewed next to traffic and engagement. We focus on fixes that matter to visitors, and we do not promise that faster pages alone will move rankings. You can see what is included on our SEO services page.
Frequently asked questions
Is page speed a Google ranking factor?
Yes, in a limited way. Google says Core Web Vitals are used by its ranking systems, but relevance comes first. A fast page with weak content will not outrank a slower page that answers the query much better.
What is a good PageSpeed Insights score?
Lighthouse scores of 90 to 100 count as good, 50 to 89 need improvement and 0 to 49 are poor. That lab score is useful for debugging, but Core Web Vitals field data is the better measure of what visitors experience.
Why are my mobile and desktop speeds so different?
Mobile tests simulate a mid-range phone on a slower network, and real mobile visitors often have less processing power. Heavy images and JavaScript hurt more on those devices. Google measures mobile and desktop separately, so check both.
How long does it take for speed fixes to show up?
Lab tools show changes right away. Field data in PageSpeed Insights covers the previous 28 days, so improvements appear gradually as new visits replace old ones.