Website Speed Statistics UK 2026

Statistics 21–30 — Mobile Performance

Mobile performance remains one of the most important challenges in modern website delivery.

Responsive design has made it possible for the same website to work across multiple screen sizes, but responsive layout alone does not guarantee a fast mobile experience. Mobile devices frequently operate with less processing power, more variable network conditions and tighter memory constraints than desktop computers.

At the same time, modern mobile pages continue to load substantial quantities of images, JavaScript, fonts and third-party resources.

The result is an important distinction: a website may look mobile-friendly while still delivering a technically demanding mobile experience.

21. Only 55% of mobile websites achieve a good First Contentful Paint

This represented an improvement from 51% in 2024, while the proportion of mobile pages with poor FCP declined from 18% to 16%.

First Contentful Paint measures how quickly the browser displays the first visible piece of page content after navigation begins.

It therefore contributes strongly to the user’s initial perception of whether the website is responding.

A page may ultimately achieve an acceptable LCP while still feeling slow if the screen remains visually empty for too long at the beginning of the experience.

Mobile implication: Early visual feedback matters because users begin judging performance before the largest page element has finished rendering.

Source: HTTP Archive, Web Almanac 2025 — Performance.

22. Only 44% of mobile websites achieve a good Time to First Byte

A further 40% fell into the “needs improvement” range, while 17% were classified as poor.

Mobile TTFB improved slightly from 42% achieving a good result in 2024, but the figures show that server and network response remain significant performance constraints.

TTFB can be affected by hosting infrastructure, redirects, network latency, backend processing, caching and content-delivery architecture.

Infrastructure implication: Front-end optimisation cannot fully compensate for consistently slow server response. Performance work needs to consider the complete delivery path.

Source: HTTP Archive, Web Almanac 2025 — Performance.

23. Only 15% of mobile pages pass the render-blocking resources audit

The rate improved only slightly from 14% in 2024.

Render-blocking stylesheets, scripts and other resources can delay the browser from displaying useful content even when the underlying network connection is reasonably fast.

This means perceived speed depends not only on how many bytes are transferred but also on the order in which critical resources are discovered, downloaded and processed.

Rendering implication: Performance optimisation should prioritise the critical rendering path rather than simply attempting to reduce total page size indiscriminately.

Source: HTTP Archive, Web Almanac 2025 — Performance.

24. The median mobile homepage now weighs approximately 2.56 MB

HTTP Archive reports that the median mobile homepage increased from approximately 2.4 MB in 2024, representing year-on-year growth of 8.4%.

The growth is striking because mobile devices are often the environments in which bandwidth and processing resources are most constrained.

A heavier page does not automatically mean a slow page, because compression, caching, prioritisation and network conditions all affect delivery. However, greater weight increases the amount of data the browser may need to transfer and process.

Page-weight implication: Mobile performance budgets remain valuable even as network speeds and device capabilities improve.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

25. Median mobile inner-page weight has increased 27.8% since 2022

This is important because users frequently enter websites directly through product pages, service pages, articles and landing pages rather than through the homepage.

Performance programmes focused only on the homepage can therefore overlook the pages that organic search, paid campaigns and AI referrals actually deliver users into.

Journey implication: Performance budgets should apply across templates and content types, not solely to the homepage.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

26. Images account for approximately 911 KB on the median mobile homepage

Images represented the largest individual content category by transferred bytes on the median mobile homepage.

By comparison, JavaScript accounted for approximately 632 KB, fonts 122 KB, CSS 77 KB and HTML 22 KB within the homepage analysis.

Image optimisation therefore remains one of the most accessible opportunities for reducing mobile payload.

However, optimisation should include more than compression. Appropriate dimensions, responsive image delivery, modern formats and correct loading priority can all influence performance.

Image implication: Mobile users should not automatically receive desktop-sized visual assets when substantially smaller resources can satisfy the viewport.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

27. Median mobile pages transfer approximately 646 KB of JavaScript

At the 90th percentile, mobile JavaScript payload rose to approximately 1.91 MB.

JavaScript creates a different performance cost from static assets such as images because scripts must not only be transferred but also parsed, compiled and executed.

This processing cost is particularly important on mobile devices with less powerful CPUs.

JavaScript implication: The number of bytes transferred understates the true cost of JavaScript because execution creates additional CPU and main-thread work after download.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

28. The median mobile page contains approximately 251 KB of unused JavaScript

At the 90th percentile, the figure increased to approximately 931 KB.

Unused JavaScript represents code that the browser downloads and processes but does not need for the measured page experience.

This can result from broad framework bundles, plugins, tag managers, third-party integrations or functionality loaded globally even when it is not required on an individual page.

Efficiency implication: One of the most valuable JavaScript optimisations is not simply compressing code but avoiding unnecessary code delivery altogether.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

29. The median mobile page makes approximately 72 network requests

Page performance is affected by more than total transferred bytes. Every individual resource request can introduce connection, scheduling and processing overhead.

HTTP Archive’s median mobile page included requests for HTML, CSS, fonts, images and a substantial number of JavaScript files.

This means a relatively lightweight page can still become inefficient if it requires a highly fragmented sequence of resource requests before becoming usable.

Request implication: Performance audits should consider both page weight and request complexity.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

30. 87% of mobile pages use at least one web font

Custom typography can strengthen visual identity, but fonts introduce additional network and rendering considerations.

If essential fonts arrive late, users may initially see fallback text before the intended font replaces it. Differences in character dimensions can then contribute to visual movement.

Despite widespread font usage, HTTP Archive found that only around 16% of mobile pages used preload for fonts in 2025.

Other resource hints were also relatively limited, with approximately 22% using preconnect and 24% using DNS prefetch.

Typography implication: Brand typography should be delivered efficiently. Font selection, subset size, preload strategy and fallback matching can all influence mobile performance.

Source: HTTP Archive, Web Almanac 2025 — Performance.

Statistics 31–40 — Conversion, Ecommerce & Revenue Impact

Technical performance becomes commercially important when improvements in loading, interaction and navigation translate into measurable changes in customer behaviour.

The relationship is not identical for every organisation. Conversion rates are affected by traffic quality, pricing, product demand, design, brand trust, seasonality and many other factors. Website speed should therefore not be treated as a guaranteed revenue formula.

However, controlled experiments and real-world performance programmes repeatedly show that reducing friction can coincide with meaningful improvements in conversion, engagement, order progression and revenue.

31. Ray-Ban increased mobile product-page conversion rates by 101.47%

The programme used the Speculation Rules API to prepare product pages before users completed the navigation.

Mobile LCP fell from 4.69 seconds to 2.66 seconds, representing an improvement of approximately 43%.

Ray-Ban compared users receiving conventionally served product pages with users receiving prerendered pages and observed substantial differences in business performance.

Ecommerce implication: Improving the transition between product-listing and product-detail pages can influence the performance of a critical part of the ecommerce journey.

Source: Google web.dev, Ray-Ban Speculation Rules Case Study, 2025.

32. Ray-Ban increased desktop product-page conversion rates by 156.16%

Desktop LCP improved from 3.03 seconds to 1.74 seconds after prerendering was introduced.

The scale of improvement should be understood within the specific experiment rather than treated as a general prediction for other ecommerce websites.

Nevertheless, the case study demonstrates that navigation performance between commercially important pages can be as significant as the initial landing-page load.

Journey implication: Ecommerce performance optimisation should consider repeat navigations between product categories, product pages, baskets and checkout rather than focusing exclusively on first load.

Source: Google web.dev, Ray-Ban Speculation Rules Case Study, 2025.

33. Ray-Ban reduced product-page exit rates by approximately 13%

Average pages viewed per session also increased by 51.95% on mobile and 65.30% on desktop.

This is important because performance can affect more than final conversion. A smoother experience can influence whether users continue exploring products rather than abandoning the journey.

Engagement implication: Exit rate, pages per session and journey progression can provide valuable supporting indicators when evaluating the commercial effects of performance improvements.

Source: Google web.dev, Ray-Ban Speculation Rules Case Study, 2025.

34. T-Mobile improved its visit-to-order rate by 60%

The programme produced an overall 42% reduction in Largest Contentful Paint.

T-Mobile combined real-user Web Vitals data with business analytics, allowing teams to examine how slower LCP corresponded with higher bounce rates and lower conversion rates.

The resulting data was used to make the business case for a cross-functional performance programme involving SEO, product and engineering teams.

Commercial implication: Connecting field performance data directly with conversion data can make technical optimisation easier to prioritise at executive level.

Source: Google web.dev, T-Mobile Web Performance Case Study, 2025.

35. T-Mobile reduced slow-loading website complaints by 34%

Overall website complaints fell by 20% during the same broader performance programme.

This illustrates a dimension of website performance that conversion analysis alone can miss: slow experiences can create service issues and customer dissatisfaction.

Performance therefore has potential implications for customer experience teams as well as marketing and ecommerce teams.

Customer-experience implication: User complaints can provide useful qualitative evidence alongside Core Web Vitals and conversion metrics.

Source: Google web.dev, T-Mobile Web Performance Case Study, 2025.

36. QuintoAndar increased conversions by 36% after major INP improvement

For QuintoAndar, a conversion represented a new lead scheduling a property visit.

The company emphasised that the increase was strongly associated with, but not solely caused by, improved user engagement resulting from the performance work.

This qualification is important because commercial metrics can change for many reasons during the same period.

Measurement implication: Performance programmes should identify associations with business results while avoiding unsupported claims that a single technical change caused every observed commercial improvement.

Source: Google web.dev, QuintoAndar INP Case Study, 2025.

37. QuintoAndar increased the proportion of pages with good INP from 42% to 78%

At the same time, the proportion delivering a poor INP experience fell from 32% to 6.9%.

The programme included real-user monitoring, removal of unnecessary third-party pixels, long-task optimisation and changes to React rendering behaviour.

This reinforces the importance of runtime responsiveness for complex digital products where users actively search, filter, compare and submit information.

Interaction implication: Commercial websites should measure the responsiveness of real interactions rather than concentrating exclusively on how quickly the first screen becomes visible.

Source: Google web.dev, QuintoAndar INP Case Study, 2025.

38. Rakuten 24 increased revenue per visitor by 53.37%

The two versions were designed to be visually and functionally identical, while the optimised version delivered stronger Web Vitals and finished loading approximately 0.4 seconds earlier in the mobile test.

The experiment provides a useful example of how performance can be tested while attempting to minimise major design and functional differences between variants.

Revenue implication: Revenue per visitor can be a valuable performance KPI because it connects technical improvements directly with the economic value of site traffic.

Source: Google web.dev, Rakuten 24 Core Web Vitals Case Study.

39. Rakuten 24 increased conversion rate by 33.13%

Average order value increased by 15.20%, average time spent increased by 9.99%, and exit rate fell by 35.12%.

The wider set of outcomes demonstrates why website performance should not be judged using a single commercial metric.

Improvements can affect several parts of customer behaviour simultaneously, including conversion, order value, engagement and abandonment.

Analytics implication: Performance programmes should track a basket of relevant business KPIs rather than searching for one universal speed-to-revenue metric.

Source: Google web.dev, Rakuten 24 Core Web Vitals Case Study.

40. Swappie increased mobile revenue by 42%

The refurbished-phone retailer tracked mobile conversion performance relative to desktop in an attempt to reduce the effect of shared campaigns and seasonality.

Its relative mobile conversion rate increased from 24% to 34% during the performance programme.

This approach is useful because it recognises that conversion rates are influenced by factors outside website speed. Comparing related user groups can sometimes provide more meaningful evidence than reviewing one headline conversion rate in isolation.

Measurement implication: Organisations should design performance measurement around the structure of their own business rather than relying entirely on generic industry benchmarks.

Source: Google web.dev, Swappie Core Web Vitals Case Study.

Statistics 41–50 — Page Weight, SEO & Modern Web Delivery

Modern website performance is increasingly shaped by the amount of data browsers need to download, the number of resources they need to request and the amount of code they must process before and after a page becomes visible.

Faster networks and more powerful devices have improved what websites can deliver, but the web has simultaneously become substantially heavier and more complex.

Images, JavaScript, video, advertising systems, analytics, consent platforms, personalisation, fonts and third-party integrations can all contribute to page weight and processing overhead.

The final ten statistics therefore examine the underlying delivery architecture behind website speed, together with the relationship between technical performance and search experience.

41. The median web page now weighs approximately 2.41 MB on desktop and 2.16 MB on mobile

These figures represent the 50th percentile across the wider dataset rather than homepages alone.

The difference between desktop and mobile has narrowed considerably over time, meaning mobile users frequently receive almost as much data as desktop users despite operating on smaller screens and, in many cases, less powerful devices.

Page weight alone does not determine performance, but every additional resource increases the potential network and processing work required to render the experience.

Delivery implication: Mobile should not automatically inherit desktop resource payloads simply because responsive layouts can display them.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

42. At the 90th percentile, pages exceed 9 MB on desktop and 8 MB on mobile

These are substantial payloads for ordinary web browsing.

Heavy pages can be particularly problematic for users on constrained mobile networks, older devices or limited data plans.

The effect is not limited to initial transfer time. Browsers may also need to decode images, parse styles, compile JavaScript and allocate memory for the downloaded resources.

Accessibility implication: Excessive page weight can create a disproportionately poor experience for users with slower hardware or connectivity.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

43. The median homepage is approximately 45% heavier than the median inner page

Homepages frequently contain large hero assets, promotional components, animations, sliders, video, tracking systems and multiple pathways into deeper areas of a website.

That complexity can make the homepage one of the heaviest templates even though many users arrive directly on inner pages through search engines, advertising campaigns, external links and AI-driven discovery.

Architecture implication: Performance optimisation should prioritise templates according to both page complexity and actual traffic-entry patterns.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

44. Mobile homepage weight has increased by approximately 203% over a decade

Desktop homepage weight increased by approximately 110.2% over the same decade.

The increase reflects the expanding capabilities of modern websites, but it also demonstrates how improvements in networks and hardware can be absorbed by additional functionality and richer media.

In other words, the web can become faster technologically while individual pages simultaneously become heavier.

Long-term implication: Performance gains from improved infrastructure can be lost when page complexity grows faster than delivery efficiency.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

45. 98.1% of web pages request at least one JavaScript file

The figure does not include JavaScript embedded directly inside HTML, meaning actual JavaScript usage is even broader.

The median desktop page made 23 JavaScript requests and the median mobile page made 22.

At the 90th percentile, those figures increased to 67 requests on desktop and 65 on mobile.

JavaScript is particularly important because it creates what HTTP Archive describes as a double performance cost: data must first be transferred and then processed by the user’s device.

Development implication: JavaScript should be evaluated according to whether its functionality justifies both its network and execution costs.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

46. 83% of LCP images still use JPG or PNG

WebP represented approximately 11% of LCP images, while AVIF remained below 1%.

This matters because the LCP image is frequently one of the most performance-sensitive resources on the page.

Modern formats can often provide comparable visual quality at lower transfer sizes, although the optimal format depends on the image characteristics, encoding settings and delivery pipeline.

Image implication: Updating image-generation and delivery workflows can create performance improvements without necessarily reducing visual quality.

Source: HTTP Archive, Web Almanac 2025 — Performance.

47. WebP gained around four percentage points of LCP-image usage in one year

AVIF reached approximately 0.7% of LCP images.

The direction shows gradual migration towards newer formats, but the pace remains relatively slow considering their broad browser support.

Legacy media libraries, CMS defaults, existing upload workflows and compatibility fallbacks can all slow adoption.

Modernisation implication: Image performance is partly an editorial and infrastructure issue; efficient formats need to be incorporated into the publishing workflow rather than applied manually page by page.

Source: HTTP Archive, Web Almanac 2025 — Performance.

48. Mobile video resources reach almost 4.8 MB at the 90th percentile

At the median, pages containing video transferred approximately 384 KB of video on mobile and 247 KB on desktop.

Median video bytes increased by 28% year on year, from approximately 246 KB in 2024 to 315 KB in 2025 across the analysed dataset.

Video can create highly effective visual experiences, but autoplay backgrounds, looping hero media and embedded players can introduce substantial payload.

Media implication: Video should be treated as a high-cost performance asset and loaded according to actual user value rather than decorative convention.

Source: HTTP Archive, Web Almanac 2025 — Page Weight.

49. The top 1,000 websites make a median 106 third-party requests on mobile

Across the wider dataset, the equivalent median was 79 requests on mobile and 83 on desktop.

Third-party requests increased year on year: the top 1,000 sites added approximately 15 requests on both desktop and mobile, while the wider dataset added around five.

These resources can include advertising, analytics, tag management, consent systems, social features, video platforms and other external services.

Governance implication: Third-party technology should be treated as part of performance governance because external scripts can create significant resource and main-thread costs.

Source: HTTP Archive, Web Almanac 2025 — Third Parties.

50. Zstandard compression adoption on CDNs increased from 3% to 12% in one year

Modern delivery networks are increasingly adopting more efficient compression technologies, although gzip remains widely used.

For origin-served resources in the analysed mobile dataset, approximately 61% used gzip compared with 39% using Brotli.

Compression is particularly useful for text-based resources including HTML, CSS, JavaScript, JSON and SVG, where substantial reductions in transfer size can often be achieved without changing the content delivered to users.

Infrastructure implication: Modern web performance depends not only on reducing the amount of content a website uses, but also on delivering necessary resources as efficiently as possible.

Source: HTTP Archive, Web Almanac 2025 — CDN.

Website Speed and SEO: What Google Actually Says

Website performance is relevant to search, but the relationship should be described accurately.

Google states that its core ranking systems use Core Web Vitals as part of the signals associated with page experience.

However, Google explicitly states that there is no single page experience ranking signal and that obtaining good Core Web Vitals scores does not guarantee that a page will rank at the top of search results.

Google also advises site owners not to pursue a perfect performance score purely for SEO reasons. Relevance and useful content remain fundamental, while strong page experience can contribute to success where multiple pages offer similarly useful information.

CGO Media Analysis — What the 50 Website Speed Statistics Tell Us

Taken together, the 50 statistics in this report show that website performance should no longer be treated as a narrow technical issue owned exclusively by developers.

Performance affects how quickly users see useful information, how easily they interact with a website, how smoothly they move through commercial journeys and how reliably pages behave across different devices and network conditions.

The evidence also demonstrates that website speed is becoming more complex. Faster hosting or image compression alone cannot solve every performance problem. Modern websites increasingly depend on JavaScript, third-party services, media assets, personalisation systems, analytics platforms and dynamic interfaces that all contribute to the final user experience.

1. Mobile Remains the Principal Performance Challenge

The strongest recurring pattern across the data is the difference between mobile and desktop performance.

Mobile users generally operate with less processing power, more variable connectivity and tighter resource constraints. At the same time, mobile pages frequently receive almost the same amount of data and functionality as their desktop equivalents.

This creates a structural performance problem.

A website can appear perfectly responsive from a design perspective while still requiring a mobile device to download and process several megabytes of resources and execute substantial amounts of JavaScript.

2. Core Web Vitals Are Useful Because They Separate Different Types of Performance Failure

A single page-load-time figure cannot explain every problem users encounter.

A website can display its main content quickly but respond slowly to interaction. It can load quickly but shift visually while the user tries to click. It can also appear responsive after a long period in which the browser initially shows very little useful information.

Core Web Vitals help separate these different experiences through loading performance, interaction responsiveness and visual stability.

This makes them valuable diagnostic indicators rather than merely search-engine metrics.

3. Website Performance Is a Journey, Not a Homepage Score

One of the most important findings from the research is the difference between homepages and deeper pages.

Users increasingly enter websites through service pages, product pages, articles, category pages, campaign landing pages and other internal URLs.

Search engines, advertising campaigns, external links and AI-assisted discovery can all send users directly into these deeper parts of the site.

A fast homepage therefore provides little protection if the pages supporting the customer journey perform poorly.

4. Images Remain One of the Largest Performance Opportunities

Images continue to dominate important parts of the loading experience.

They frequently represent the Largest Contentful Paint element and remain one of the largest contributors to transferred page weight.

This gives organisations several practical optimisation opportunities:

• Appropriate image dimensions.
• Responsive image delivery.
• Modern formats such as WebP and AVIF where appropriate.
• Compression.
• Correct loading priority.
• Avoidance of unnecessary high-resolution assets.
• Removal of decorative images that provide limited value.

The issue is particularly relevant to websites with large hero imagery, product photography, sector graphics, featured images or background media.

5. JavaScript Is Becoming One of the Biggest Sources of Performance Risk

JavaScript is now effectively universal across the modern web.

Its performance cost is different from many other resources because browsers must do more than download it. JavaScript needs to be parsed, compiled and executed, often while other work is competing for the same main thread.

This matters disproportionately on mobile devices.

Large frameworks, plugins, trackers, widgets and tag-management systems can all increase the amount of work required before a page becomes fully responsive.

The research also shows substantial quantities of unused JavaScript being delivered to users.

6. Third-Party Technology Needs Performance Governance

Modern websites depend heavily on third-party services.

Analytics, advertising, consent tools, video platforms, live chat, social widgets, payment technology, personalisation systems and marketing automation can all provide legitimate business value.

But they can also create substantial network requests and additional JavaScript execution.

The performance cost of these tools can become difficult to control because the organisation does not fully own the code being delivered.

7. Faster Does Not Mean Stripping Functionality From the Website

Performance optimisation is sometimes presented as a choice between rich functionality and speed.

That is an oversimplification.

The stronger approach is to determine which resources and features genuinely contribute to customer value and then deliver those features efficiently.

Modern techniques such as lazy loading, prioritisation, caching, code splitting, prerendering, compression and responsive media can allow websites to retain useful functionality while reducing unnecessary work.

8. The Commercial Value of Speed Depends on the Journey

The case studies analysed in this report show substantial commercial improvements following performance programmes, but the size of those improvements varies considerably.

This is expected.

A faster ecommerce product page, a responsive property-search interface and a quicker telecommunications checkout are different user experiences with different commercial dynamics.

The important principle is therefore not that every 100-millisecond improvement creates an identical financial return.

The stronger conclusion is that technical friction can influence user behaviour, and organisations should measure that relationship using their own data.

9. Website Speed Is Not an SEO Shortcut

Performance matters to search experience, but it should not be exaggerated into a claim that faster websites automatically outrank slower competitors.

Google uses Core Web Vitals within its wider systems associated with page experience, but performance exists alongside relevance, usefulness, authority and many other factors.

A page with excellent performance but weak content will not become a strong search result purely because it loads quickly.

Equally, a highly relevant and authoritative page can still provide a poor user experience when performance is neglected.

10. Performance Needs Continuous Management

A website can be fast when launched and become progressively slower over time.

New plugins, advertising scripts, tracking systems, images, videos, campaigns, page-builder elements and functionality can gradually increase weight and complexity.

This makes performance an operational issue rather than a one-time technical project.

Regular monitoring can identify regressions before they become embedded across large parts of the website.

What UK Businesses Should Do in 2026

The evidence in this report shows that website performance should be managed as an ongoing business discipline rather than a one-off technical project.

For UK organisations, the objective should not be to chase perfect laboratory scores. It should be to identify the pages and journeys that matter most to users, remove unnecessary performance friction and create a process that prevents performance from deteriorating as the website evolves.

1. Measure Real Users, Not Just Laboratory Tests

Laboratory tools such as Lighthouse and PageSpeed Insights are useful for diagnosis, but they do not replace real-user data.

Businesses should monitor how actual visitors experience the site across different devices, connection speeds and page types.

Priority metrics should include:

• Largest Contentful Paint.
• Interaction to Next Paint.
• Cumulative Layout Shift.
• Time to First Byte.
• First Contentful Paint.
• Conversion rate.
• Exit rate.
• Lead-generation completion.
• Ecommerce journey progression.

2. Prioritise Mobile Performance First

Mobile remains the weaker environment across several important performance measures.

This means optimisation should be tested on realistic mobile devices rather than powerful desktop hardware alone.

Areas to review include:

• Mobile LCP.
• Main-thread blocking.
• JavaScript execution.
• Image dimensions.
• Font loading.
• Third-party scripts.
• Menu and form responsiveness.
• Product filters and interactive components.

3. Optimise Images as Part of the Publishing Workflow

Images remain one of the largest contributors to page weight and frequently determine Largest Contentful Paint.

Image optimisation should therefore be built into the publishing process rather than performed manually after pages become slow.

Businesses should consider:

• WebP or AVIF where appropriate.
• Correct image dimensions.
• Responsive image delivery.
• Compression.
• Lazy loading below the fold.
• Prioritisation of LCP images.
• Avoidance of unnecessarily large featured images.
• Removal of duplicate image variants from the same viewport.

4. Control JavaScript Aggressively

JavaScript is now present on almost every modern website, but it remains one of the most expensive resources from a performance perspective.

Businesses should identify which scripts are genuinely required and which are being loaded by default because of old plugins, abandoned experiments or global theme settings.

Common areas to investigate include:

• Page-builder scripts.
• Unused plugins.
• Tag-manager containers.
• Live chat.
• Social widgets.
• Advertising technology.
• Analytics duplication.
• Animation libraries.
• Third-party forms.

5. Introduce Performance Budgets

Without explicit limits, websites often become progressively heavier over time.

A performance budget establishes acceptable thresholds for metrics such as:

• Total page weight.
• Image payload.
• JavaScript payload.
• Number of third-party requests.
• LCP.
• INP.
• CLS.
• Server response time.

These budgets can be applied to new templates, redesigns, campaigns and major development releases.

6. Review Third-Party Technology Regularly

Third-party scripts can accumulate quickly because they are often added by multiple departments.

Marketing, analytics, advertising, compliance, sales and customer-service teams may all add their own technologies without visibility into the total performance cost.

This can result in duplicated or low-value scripts being loaded across the entire site.

7. Optimise the Pages That Generate Revenue First

Not every page carries the same commercial importance.

Performance work should therefore prioritise areas such as:

• Product pages.
• Category pages.
• Service pages.
• Lead-generation pages.
• Checkout.
• Application forms.
• Booking journeys.
• Pricing pages.
• High-traffic organic landing pages.

This allows technical investment to be connected directly with commercial outcomes.

8. Improve Hosting and Server Response Where Necessary

Front-end improvements cannot fully compensate for consistently slow server response.

Businesses should review:

• Hosting capacity.
• Server location.
• Full-page caching.
• Object caching.
• Database performance.
• CDN configuration.
• Redirect chains.
• Backend processing.
• Compression.

The objective is to reduce the delay before the browser receives useful page data.

9. Treat WordPress and Page Builders as Performance Systems

WordPress websites can perform extremely well, but themes, page builders and plugins introduce layers of functionality that need active management.

For Avada and similar systems, businesses should review:

• Unused global elements.
• Slider libraries.
• Animation effects.
• Large background images.
• Duplicate desktop and mobile assets.
• Unnecessary font variants.
• Global scripts loaded on pages that do not use them.
• Plugin overlap.
• Cache configuration.

A flexible page builder can make publishing faster while simultaneously creating performance overhead if every available feature is enabled without discipline.

10. Monitor Performance After Every Major Change

Performance regressions often occur gradually.

A new tracking script, larger hero image, redesigned form, additional font or plugin update can degrade performance without being immediately obvious.

Organisations should monitor performance after:

• Major content uploads.
• Website redesigns.
• Plugin installations.
• Campaign launches.
• Analytics changes.
• Theme updates.
• New third-party integrations.
• CMS upgrades.

Website Performance Trends UK 2026–2027

The evidence reviewed throughout this report suggests that website performance is moving away from isolated speed optimisation and towards broader performance governance.

Pages continue to become heavier, JavaScript remains almost universal, third-party requests continue to grow and mobile devices still face significantly greater processing constraints than desktop environments.

At the same time, browser capabilities, modern image formats, speculative navigation, caching, compression and real-user monitoring provide organisations with more tools than ever to improve performance without necessarily sacrificing functionality.

The following trends are therefore presented as evidence-led directions rather than guaranteed predictions.

Trend 1 — Real-User Performance Becomes More Important Than Isolated Lab Scores

Laboratory tests remain useful because they provide controlled diagnostics, but real users experience websites across different hardware, networks, browsers and journey stages.

This makes field data increasingly important for organisations trying to determine whether technical changes are improving actual experiences.

Real-user monitoring can reveal whether performance problems affect particular devices, geographies, templates or customer journeys in ways that a single laboratory test may not expose.

Trend 2 — Performance Budgets Become Part of Website Governance

HTTP Archive continues to record growth in page weight and request volume.

This creates a predictable operational problem: websites can become progressively slower as teams add images, plugins, analytics, personalisation, advertising and interactive functionality.

Performance budgets provide a way to manage that growth by establishing acceptable limits before new functionality is released.

Possible budgets can cover:

• Total transferred bytes.
• JavaScript payload.
• Image weight.
• Third-party requests.
• LCP.
• INP.
• CLS.

Trend 3 — JavaScript Restraint Becomes More Important

JavaScript now appears on almost every modern website and remains one of the most significant sources of runtime performance cost.

The issue is not simply file size. Scripts need to be downloaded, parsed, compiled and executed, creating additional CPU and main-thread work.

HTTP Archive also continues to identify substantial quantities of unused JavaScript across ordinary pages.

This is likely to increase pressure on development teams to use selective loading, code splitting and simpler architectures where possible.

Trend 4 — Third-Party Scripts Receive Greater Scrutiny

More than 90% of pages in HTTP Archive’s third-party research use at least one third-party service.

Advertising, analytics, tag management, consent platforms, social functionality, video and marketing technologies can all introduce additional requests and processing costs.

The number of third-party requests has continued to increase even while the number of unique third-party domains has declined.

This suggests that fewer providers can still create substantial performance overhead through larger numbers of individual requests.

Trend 5 — Image Optimisation Becomes More Automated

Images remain one of the largest contributors to transferred page weight and are frequently the Largest Contentful Paint element.

Yet legacy formats continue to dominate many important visual resources.

The gradual increase in WebP adoption demonstrates the direction of travel, while modern content-delivery systems increasingly automate resizing, compression and format selection according to browser and viewport.

This makes manual image optimisation less sustainable for large websites.

Trend 6 — Speculative Navigation and Browser Caching Become More Valuable

Traditional performance optimisation often concentrates on initial page load.

Recent case studies demonstrate growing attention to what happens after the user begins navigating through the website.

Ray-Ban used the Speculation Rules API to prerender likely future product pages, improving LCP and conversion performance. Monrif subsequently reported LCP improvements of up to 17.9% and engagement increases of up to 8.9% after introducing prerendering and back/forward cache improvements.

These techniques can create near-instant transitions where likely future navigations can be predicted safely.

Trend 7 — CMS and Page-Builder Performance Becomes a Strategic Issue

Content management systems allow organisations to publish quickly, but additional themes, page builders, components, plugins and integrations can increase page weight and JavaScript execution.

HTTP Archive’s 2025 CMS research shows substantial differences in resource weight between major platforms, with JavaScript remaining one of the largest drivers of variation.

This does not mean that a particular CMS automatically produces a slow website. Implementation decisions remain critical.

However, organisations using flexible systems such as WordPress increasingly need standard performance rules for templates, plugins, images and global components.

Trend 8 — Performance and Accessibility Become More Closely Connected

Heavy pages disproportionately affect users on slower connections and less powerful devices.

HTTP Archive explicitly highlights the relationship between increasing page weight and the digital divide.

A site that performs well on a current flagship phone over fast Wi-Fi may deliver a substantially weaker experience to users relying on older devices or constrained mobile networks.

Performance therefore has an inclusion dimension as well as a conversion and SEO dimension.

Trend 9 — AI-Assisted Performance Diagnostics Expand

Website performance produces large volumes of technical data across waterfalls, Core Web Vitals, JavaScript execution, request chains and real-user monitoring.

AI-assisted development tools can increasingly help teams interpret that information, identify probable bottlenecks and suggest areas for investigation.

However, automated recommendations still need human validation because the correct optimisation depends on the website’s architecture, business objectives and user journey.

A tool may correctly identify a large script while lacking the commercial context required to know whether that functionality is essential.

Trend 10 — Performance Remains Part of SEO, But Not an SEO Shortcut

The relationship between website performance and search visibility is likely to remain frequently misunderstood.

Google continues to use Core Web Vitals within its wider systems associated with page experience, but it explicitly states that good Core Web Vitals alone do not guarantee high rankings.

The practical effect is that performance remains important without replacing relevance, usefulness, authority or content quality.

This is likely to push technical SEO teams towards a more balanced model in which performance is treated as one component of overall search quality rather than an isolated ranking tactic.

Research Methodology & Limitations

This report was developed by the CGO Media Research Team to provide a structured overview of current website speed, Core Web Vitals, mobile performance, page-weight and conversion evidence relevant to UK organisations in 2026.

The objective was not simply to assemble a large collection of frequently quoted website-speed claims. Priority was given to statistics with identifiable sources, documented methodologies and sufficient context to understand what the figures actually represent.

Research Scope

The report focuses on five connected areas of website performance:

• Website speed and user behaviour.
• Core Web Vitals and technical performance.
• Mobile performance.
• Conversion, ecommerce and revenue impact.
• Page weight, SEO and modern web delivery.

The statistics were selected to help explain not only how fast modern websites are, but also why performance varies and how technical improvements can influence user and commercial outcomes.

Primary Data Sources

CGO Media prioritised primary and large-scale research sources wherever possible.

Principal sources used within the report include:

• HTTP Archive.
• The Web Almanac.
• Chrome UX Report data.
• Google web.dev case studies.
• Deloitte performance research.
• Portent conversion research.
• Google Search documentation.

Where statistics originated from individual company case studies, those findings are presented as evidence from a specific environment rather than universal benchmarks.

Field Data and Laboratory Data

Website-performance research frequently uses two different forms of measurement: field data and laboratory data.

Field data reflects the experiences of real users across different devices, networks and environments. Examples include Chrome UX Report measurements of Core Web Vitals.

Laboratory data measures performance under controlled test conditions. These tests can help diagnose technical causes but may not represent every real-world user experience.

Both forms of evidence are useful, but they answer different questions.

Core Web Vitals Measurement

Core Web Vitals are assessed using the performance thresholds defined by Google for loading, responsiveness and visual stability.

The report therefore distinguishes between:

• Largest Contentful Paint.
• Interaction to Next Paint.
• Cumulative Layout Shift.
• Supporting metrics such as First Contentful Paint and Time to First Byte.

The overall Core Web Vitals pass rate requires all three Core Web Vitals to meet their recommended thresholds at the required percentile.

A website can therefore perform strongly on one metric while still failing the overall assessment because of weakness elsewhere.

Page Weight and Resource Analysis

Page-weight statistics refer to the amount of transferred data associated with websites or particular resource categories.

Transferred bytes are not identical to the uncompressed size of resources after they reach the browser.

JavaScript, HTML, CSS and other text resources may be compressed during transfer but create greater processing costs after decompression and execution.

For this reason, page weight should not be treated as a complete measurement of performance by itself.

Commercial Case Studies

Several statistics in this report come from individual performance programmes conducted by organisations such as ecommerce, telecommunications and property businesses.

These case studies are valuable because they connect technical changes with real commercial metrics such as:

• Conversion rate.
• Revenue per visitor.
• Order value.
• Exit rate.
• Customer complaints.
• Engagement.

However, case-study outcomes should not be generalised automatically to other businesses.

Results can vary according to traffic quality, product demand, brand strength, seasonality, pricing, website architecture and the specific performance bottleneck being addressed.

Historical Statistics

Some widely cited website-speed statistics originate from earlier research.

These figures remain useful where they illustrate important behavioural relationships, but the publication and measurement dates must be retained.

For example, historical mobile abandonment research should not be presented as if it were a newly measured UK statistic from 2026.

CGO Media therefore distinguishes between:

• Current web-performance benchmarks.
• Recent commercial case studies.
• Historical behavioural evidence.

UK-Specific and International Evidence

The report is written for UK organisations, but not every source is UK-specific.

Large-scale technical datasets such as HTTP Archive describe the wider web rather than UK websites alone.

Similarly, individual company case studies may originate from organisations operating in other markets.

These sources remain valuable for identifying performance patterns, but they should not be used to claim that an identical percentage applies specifically to all UK websites or consumers.

Correlation and Causation

Website performance and commercial results can move together without performance being the sole cause of the change.

Marketing campaigns, pricing, product availability, redesigns, seasonality and changes in traffic composition can all affect conversion during the same period.

For that reason, this report avoids treating every association between faster performance and improved revenue as proof of direct causation.

Controlled experiments and comparison groups provide stronger evidence than simple before-and-after observations, but even controlled studies need to be interpreted within their own methodology.

Website Speed and SEO

Search performance requires particularly careful interpretation.

Core Web Vitals form part of Google’s broader systems associated with page experience, but Google does not describe website speed as a standalone shortcut to higher rankings.

Pages still need to provide relevant, useful and authoritative information.

For that reason, this report separates the user-experience and commercial value of performance from exaggerated claims that achieving a particular speed score automatically produces ranking gains.

AI Search and Website Performance

AI search systems introduce new forms of digital discovery, but current public evidence does not justify claiming that faster websites receive direct AI-ranking benefits simply because they are faster.

Technical accessibility, crawlability, content structure and reliable delivery remain important elements of a well-maintained website, but these concepts should not be converted into unsupported AI-search ranking claims.

Research Limitations

No single statistics report can represent every website, industry, user or technical environment.

The principal limitations of this research include:

• Not all evidence is UK-specific.
• Some commercial statistics come from individual company case studies.
• Historical behavioural studies may not reflect current user behaviour exactly.
• Performance varies significantly between devices and network conditions.
• Laboratory testing does not reproduce every real-world environment.
• Commercial outcomes are influenced by factors beyond website speed.
• Core Web Vitals and browser technologies can evolve.
• Large-scale datasets represent aggregated patterns rather than every individual website.
• Page weight does not by itself explain perceived speed.
• New research may supersede individual findings after publication.

These limitations define how the evidence should be interpreted rather than removing its value.

Frequently Asked Questions — Website Speed Statistics UK 2026

The questions below address some of the most common issues raised by the website-speed, Core Web Vitals and commercial-performance evidence reviewed in this report.

They are designed to help businesses, marketers, developers and researchers distinguish between genuine performance evidence and oversimplified claims about page speed.

1. What is considered a fast website in 2026?

There is no single load-time number that defines every website as fast or slow.

A modern performance assessment should consider how quickly useful content appears, how responsive the page is when users interact with it and whether the layout remains stable.

Google’s current Core Web Vitals thresholds define a good experience as:

• LCP: 2.5 seconds or less.
• INP: 200 milliseconds or less.
• CLS: 0.1 or less.

2. Does website speed affect SEO rankings?

Website performance can contribute to Google’s wider page-experience systems, but speed is not a standalone shortcut to high rankings.

Google explicitly states that good Core Web Vitals do not guarantee top search positions.

Relevance, usefulness, authority and content quality remain fundamental.

3. What are Core Web Vitals?

Core Web Vitals are Google’s current set of user-experience metrics focused on three areas:

• Largest Contentful Paint — loading performance.
• Interaction to Next Paint — interaction responsiveness.
• Cumulative Layout Shift — visual stability.

They are based on how real users experience webpages rather than simply how quickly a controlled test can download resources.

4. Why does mobile performance usually lag behind desktop?

Mobile devices generally have less processing power and operate across more variable network conditions than desktop computers.

At the same time, many websites deliver similar quantities of JavaScript, imagery and third-party functionality to both environments.

This creates a larger performance burden on mobile hardware.

5. How much does page weight matter?

Page weight matters because larger resources take more data to transfer and may require more work for the browser to process.

However, page weight alone does not determine perceived speed.

A smaller page can still perform poorly if it contains render-blocking resources or inefficient JavaScript, while a heavier page may perform reasonably well if resources are prioritised, compressed and cached effectively.

6. What is the biggest cause of slow websites?

There is no universal single cause.

Common performance bottlenecks include:

• Slow server response.
• Oversized images.
• Large JavaScript bundles.
• Third-party scripts.
• Render-blocking CSS.
• Web fonts.
• Video.
• Too many network requests.
• Heavy page-builder output.
• Poor caching or CDN configuration.

The correct solution depends on which bottleneck is actually affecting the site.

7. Are images still important for website speed?

Yes.

Images remain one of the largest contributors to page weight and frequently determine Largest Contentful Paint.

Performance can often be improved through:

• WebP or AVIF.
• Correct dimensions.
• Responsive images.
• Compression.
• Lazy loading.
• Prioritisation of the LCP image.

8. Why is JavaScript such a performance issue?

JavaScript creates more than a download cost.

The browser must also parse, compile and execute it, often on the main thread.

Large or unnecessary scripts can therefore delay interaction even after the page looks visually complete.

This effect is particularly important on mobile devices with weaker CPUs.

9. Can a faster website improve conversion rates?

Yes, performance improvements have been associated with better conversion, revenue and engagement in numerous case studies.

However, the size of the effect varies substantially between businesses and customer journeys.

A technical improvement does not guarantee the same commercial uplift on every website.

10. Is a PageSpeed Insights score of 100 necessary?

No.

A perfect laboratory score is not required for strong SEO, strong conversions or good user experience.

The aim should be to identify meaningful performance problems and improve real-user outcomes rather than optimise solely for a headline number.

11. How often should website speed be checked?

Performance should be monitored continuously where possible and reviewed after major site changes.

Particular attention should be given after:

• New plugin installations.
• Theme changes.
• Large content uploads.
• Tracking changes.
• Campaign launches.
• Third-party integrations.
• Design changes.
• CMS updates.

12. What should a business optimise first?

The first priority should be the bottlenecks affecting commercially important user journeys.

A useful sequence is:

1. Measure real-user performance.
2. Identify high-value landing and conversion pages.
3. Find the dominant performance bottleneck.
4. Reduce unnecessary resource cost.
5. Improve server and delivery efficiency.
6. Optimise images and JavaScript.
7. Retest user and commercial outcomes.
8. Introduce monitoring to prevent regression.

Conclusion — Website Speed Statistics UK 2026

The evidence reviewed throughout this report shows that website performance remains a material part of digital experience in 2026.

The web has become more capable, but also more complex. Pages are heavier, JavaScript is almost universal, third-party systems are widespread and mobile users still face greater technical constraints than desktop users.

At the same time, browsers, hosting infrastructure, image formats, caching, compression and real-user monitoring have improved significantly.

The result is not a simple story of websites becoming either faster or slower. Performance increasingly depends on how effectively organisations manage complexity.

Performance Is a User Experience Issue

Core Web Vitals have helped move the performance discussion away from one-dimensional load-time metrics.

Users experience performance through several different stages:

• How quickly the first useful content appears.
• How quickly the main content becomes visible.
• How rapidly the page responds to interaction.
• Whether the layout remains stable.
• How quickly subsequent pages load.

A website can therefore be visually attractive and technically functional while still delivering a frustrating experience.

Mobile Performance Remains the Greater Challenge

The research consistently shows weaker mobile performance across several important measures.

This matters because mobile users frequently receive almost desktop-level resource complexity while operating on less powerful hardware and more variable networks.

Responsive design alone is therefore not enough.

Organisations need to consider whether mobile users are receiving appropriately sized images, efficient scripts, prioritised resources and interactions that remain responsive on realistic devices.

Page Complexity Is Increasing

The growth in page weight, JavaScript and third-party resources demonstrates why performance cannot be solved once and then ignored.

Websites naturally accumulate additional functionality over time.

Marketing tools, analytics systems, consent platforms, videos, design effects, plugins and tracking technologies can each appear relatively small when added individually while creating substantial aggregate cost.

The challenge is therefore increasingly one of governance rather than emergency repair.

Commercial Performance Should Be Measured Directly

The case studies reviewed in this report show meaningful improvements in conversion, engagement and revenue alongside website-performance programmes.

However, the size of those effects varies significantly.

No responsible analysis should therefore promise that reducing load time by a fixed amount will generate the same revenue improvement for every organisation.

Businesses should instead connect technical performance with their own conversion and journey data.

That can include:

• Conversion rate.
• Revenue per visitor.
• Order value.
• Lead completion.
• Exit rate.
• Cart engagement.
• Customer complaints.

Speed Supports SEO, But Does Not Replace It

Website performance contributes to the broader quality of the search experience, but it should not be exaggerated into a standalone ranking strategy.

Good Core Web Vitals do not guarantee strong rankings.

A technically excellent page still needs relevant, useful and authoritative content.

The strongest interpretation is therefore that performance supports SEO by improving the overall experience available to users after discovery.

Performance Is Becoming an Operating Discipline

The wider conclusion from the research is that website performance increasingly needs to be managed continuously.

A modern performance programme includes:

• Real-user monitoring.
• Laboratory diagnostics.
• Performance budgets.
• Image standards.
• JavaScript control.
• Third-party governance.
• Infrastructure review.
• Commercial measurement.
• Regression monitoring.

This changes the role of website speed from an occasional optimisation project into an ongoing part of website management.

Research Usage, Citation & Press

CGO Media publishes research, statistics, frameworks and technical analysis to support businesses, journalists, researchers and organisations examining the changing relationship between website performance, search visibility, digital experience and commercial outcomes.

The statistics and analysis within this report may be referenced in editorial coverage, research papers, presentations, technical documentation and industry commentary, provided appropriate attribution and research context are retained.

Using Individual Statistics

Where an individual statistic originates from an external organisation, CGO Media recommends citing the original research as the primary source whenever possible.

CGO Media may be cited for the synthesis, interpretation and UK-focused analysis surrounding that statistic.

This helps preserve the distinction between source evidence and CGO Media’s own research commentary.

Using CGO Media Analysis

The analytical sections of this report — including the interpretation of the 50 statistics, implications for UK organisations, performance-priority model and 2026–2027 trends analysis — represent CGO Media Research Team analysis.

Journalists, publishers and researchers are welcome to quote or summarise these findings with attribution to:

Press & Media Enquiries

Journalists, editors, broadcasters, researchers and technical publications requiring additional commentary, clarification or access to CGO Media research can use the dedicated Press & Media area.

Relevant enquiry topics include:

• Website speed and Core Web Vitals.
• Technical SEO.
• Mobile performance.
• Page weight and JavaScript.
• Conversion and ecommerce performance.
• Search visibility.
• AI search and technical website quality.
• Digital experience and performance governance.

Research Methodology

CGO Media publishes its wider research methodology separately to explain how evidence is selected, assessed and interpreted across the Research Library.

The methodology covers areas including source selection, primary and secondary evidence, international datasets, methodological limitations, AI-search evidence and responsible interpretation.

View the CGO Media Research Methodology

Research Updates

Website performance technologies, browser behaviour, Core Web Vitals and web-delivery practices continue to evolve.

CGO Media therefore reviews its technical research periodically as new datasets, browser capabilities and large-scale performance studies become available.

Readers should check publication and update information when citing individual figures, particularly where the underlying research involves fast-moving technologies or performance standards.

Editorial & Research Attribution

This report was prepared and reviewed by the CGO Media Research Team as part of CGO Media’s wider research programme covering search, technical SEO, AI search, GEO, digital authority and online discovery.

The report combines external evidence with CGO Media analysis to examine how modern website performance affects users, commercial journeys and technical search quality.

Sources & References

The Website Speed Statistics UK 2026 report draws on large-scale web-performance datasets, official Google documentation, documented company experiments and established commercial-performance research.

CGO Media has prioritised original research and primary sources wherever reasonably available. Readers, journalists and researchers are encouraged to consult the original publications when citing individual statistics.

Primary Web Performance Research

Google & Core Web Vitals Documentation

Recent Commercial Performance Case Studies

Website Speed & Conversion Research

How These Sources Were Used

The evidence in this report falls into three principal categories.

Source Interpretation Policy

CGO Media does not treat all statistics as equivalent evidence.

When interpreting the figures in this report:

• Primary sources are preferred over third-party statistics roundups.
• Research dates are retained where material.
• Global statistics are not automatically labelled as UK statistics.
• Individual company case studies are not presented as universal benchmarks.
• Correlation is distinguished from demonstrated causation.
• Google documentation is used for claims about Google Search and Core Web Vitals.
• CGO Media analysis is distinguished from the underlying source statistic.