What is Page Speed?

Page speed is not a single number. Google uses a collection of metrics grouped under the Core Web Vitals framework. These assess how fast a page feels to a real user and that distinction between feeling fast and being fast is worth dwelling on. Largest Contentful Paint (LCP) measures how long it takes for the biggest visible element to load. Cumulative Layout Shift (CLS) captures how much the page jumps around as assets load in. Interaction to Next Paint (INP) tracks how quickly the page responds when a user taps or clicks something.

Each of these metrics targets a different dimension of the user experience. LCP captures the moment a visitor feels the page has arrived. CLS measures the frustration of reaching for a button that moves before you can press it. INP reflects whether the site does anything when you interact with it. A site can score well on one and catastrophically on another, which is why chasing a single speed score misses the point. Google’s PageSpeed Insights tool draws on both lab data and real-world Chrome User Experience Report (CrUX) data. It gives you a far richer picture than any single load-time figure can.

What makes this complicated for most site owners is that speed varies by device and connection. On a MacBook Pro connected to fibre broadband, a site can load cleanly. On a mid-range Android phone on 4G, that same site can be a broken, half-rendered mess. Google’s assessments lean heavily on mobile performance and so do the people visiting your site. In the UK, mobile accounts for well over half of all web traffic. Desktop-only speed testing is a comfortable illusion.

How Page Speed Became a Serious Ranking Factor

Google announced page speed as a ranking signal for desktop searches back in 2010. The 2018 Speed Update brought it into mobile search rankings and shifted the conversation significantly. Before that, you could get away with a slow site as long as your content was strong enough. After it, technical performance became a prerequisite for competing at the top of the search results pages, not an optional extra.

The introduction of Core Web Vitals as ranking signals in 2021 sharpened this further. Sites with poor LCP, CLS or INP scores faced a measurable disadvantage in rankings, all else being equal. The phrase “all else being equal” does a lot of work there. Google has been clear that content quality still outranks page experience signals in most cases. A mediocre but fast page does not automatically outrank an excellent but slightly slower one. Speed functions more like a tiebreaker or a minimum threshold. That framing makes it more important in competitive niches where multiple strong pieces of content are fighting for the same positions.

The practical consequence is that page speed now links back to two different important factors for website SEO. It affects organic search visibility directly through rankings. It affects conversion rates separately, because users who experience a slow site bounce at higher rates regardless of how they arrived. Amazon famously found that every 100ms of additional latency cost them 1% in sales. That figure is from 2006 and the web has only become faster and more competitive since. The tolerance threshold for slow loading has moved in one direction only.

Root Technical Causes of a Slow Page

Most page speed problems trace back to a handful of root causes. Unoptimised images are the single most common culprit. A product page loading six JPEG files each weighing 2MB can be rebuilt with modern WebP or AVIF formats at a fraction of the file size. There is no perceptible loss of visual quality in doing so. Images that are not lazily loaded force the browser to fetch every image on the page at once, rather than waiting until they scroll into view. That multiplies the problem significantly.

Render-blocking resources are the second major category. JavaScript and CSS files in the document head force the browser to pause rendering until they have been downloaded and processed. Large third-party scripts, analytics tags, chat widgets and advertising pixels are frequent offenders. A site with ten different marketing tools each injecting their own JavaScript has, in effect, made a choice. Every visitor must wait while a small committee of scripts loads before anything visible appears on screen.

Server performance forms the third pillar. Time to First Byte (TTFB) measures how long the server takes to begin responding after a request. A high TTFB means every subsequent step starts from a disadvantaged position. Inefficient database queries or the absence of server-side caching can each drive TTFB to a level where front-end work alone cannot fully compensate. A Content Delivery Network (CDN) serves assets from locations closer to the visitor, but it is not a substitute for a well-configured origin server.

Core Web Vitals in Practice

Take a typical e-commerce site built on an off-the-shelf theme with a page builder and a full-screen hero video autoloading on mobile. Its LCP score is likely to be in the red. The hero video or a large promotional banner is probably the Largest Contentful Paint element and videos load slowly on mobile networks by design. Switching to a static image or deferring the video entirely for mobile users can move the LCP score from four seconds to under two. That single change can have a measurable effect on both rankings and bounce rate.

CLS problems often originate with web fonts and images that load without declared dimensions. When a browser does not know how tall an image will be, it reserves no space for it. The image then pops in and shoves the text below it downward. Users who were mid-scroll find themselves reading a different paragraph to the one they were on a second ago. Declaring explicit width and height attributes on images and using font-display for web fonts addresses most CLS issues without any visual compromise.

INP became a Core Web Vital in March 2024, replacing the older First Input Delay metric. Sites with heavy JavaScript, particularly those built on React or similar frameworks, can struggle with INP even when visual load times appear acceptable. A page that looks ready but does not respond to a tap for 400ms or more fails the INP threshold. Long JavaScript tasks block the browser’s main thread. Breaking these into smaller chunks, deferring non-essential processing or moving work to web workers are the standard remedies.

How Page Speed Scores Connect to Business Metrics

Conversion rate data from across the industry points consistently in the same direction. Portent published research showing that e-commerce conversion rates fall with each additional second of load time between zero and five seconds. Conversion rates drop by an average of 4.42% with each additional second of load time between 0 and 5 seconds. Google’s own data found that as page load time goes from one second to three seconds, the probability of a mobile user bouncing increases by 32%. The numbers vary by study and context, but the direction never does.

Organic search is where the compounding effect becomes most apparent. A page that loads slowly does not just rank slightly lower. It accumulates fewer backlinks because fewer people share content they cannot easily reach. It generates less time-on-page data, which Google’s user behaviour signals interpret as lower relevance. It earns fewer return visits. Each of these factors feeds back into rankings in ways that are hard to attribute specifically to speed but impossible to separate from it either.

For publishers and content businesses, advertising revenue adds another dimension. Programmatic ad platforms factor viewability and session duration into the prices they pay. When a significant proportion of users bail before the ads even render, the page generates less ad revenue per session than a fast one where users stay and scroll. Page speed, in that context, is literally printing money or failing to.

Tools Used to Understand Your Page Speed

Google’s PageSpeed Insights is the obvious starting point. It uses real user data from Chrome alongside lab tests and it is the same tool Google’s systems use to evaluate your site. It does give you scores for individual URLs rather than a site-wide view, which limits its usefulness for large sites with hundreds of pages. WebPageTest offers considerably more depth. You can run tests from specific locations and specify the device type and connection speed. Filmstrip views show exactly which elements load at which point in the timeline. GTmetrix sits somewhere between the two in terms of detail and accessibility.

For ongoing monitoring rather than one-off audits, tools like SpeedCurve or Calibre track your Core Web Vitals over time. They alert you when a deployment causes a regression, which is how performance problems usually appear in practice. Google Search Console’s Core Web Vitals report deserves particular attention because it surfaces real-world data for your entire site grouped by URL pattern. A site with five thousand product pages sharing the same template will show aggregated performance data for that group. This makes it far easier to prioritise fixes. A single template improvement can move hundreds of URLs from failing to passing status in one deployment.

Similarities Between Page Speed and Accessibility

Page speed and accessibility share more common ground than is immediately obvious. Both are improved by clean, semantic HTML that does not rely on JavaScript to render core content. Both benefit from smaller file sizes and reduced dependency on third-party resources. Both suffer when a site is built with visual aesthetics as the primary concern and performance as an afterthought.

Slow pages disproportionately affect users on lower-end devices and slower connections. In the UK, that correlates with older users, people in rural areas and those on lower incomes. Building a fast site also builds a more accessible one. The Web Content Accessibility Guidelines do not mandate speed targets, but the underlying principle of providing equivalent access to content applies here too.

For businesses making the case internally for page speed investment, accessibility arguments can open budget conversations that purely technical ones close. Framing a site speed project as an accessibility and inclusivity initiative, as well as an SEO and conversion one, tends to get more traction with decision-makers. They often do not immediately connect milliseconds to margins. The connection is real and making it explicit is a legitimate strategy.

With nearly two decades of experience in website design and SEO, we understand what it takes to build a site that performs well in both search rankings and real-world user experience. Fast page speed is necessary for what we do, whether that means auditing an existing site and identifying specific fixes or rebuilding a slow one from the ground up. Get in touch to find out how we can help your site load faster and rank higher.