What One Second of Load Time Really Costs Your Business

Vodafone ran an A/B test on two identical landing pages. Same design, same content, same offer. The only difference was that one had been optimised to load 31% faster. The faster page generated 8% more sales. Not 8% more traffic — 8% more completed purchases, from the same audience, seeing the same thing.

That single result, documented by Google's web.dev team, captures something that a decade of research has consistently shown: website speed isn't a technical nicety. It's a revenue variable. The BBC loses 10% of its audience for every additional second a page takes to load. More than half of all mobile websites fail Google's own performance assessment. And the businesses that have measured the relationship between load time and money have found, without exception, that faster pages sell more.

This article assembles that evidence — the landmark studies, the company-level case data, and the UK benchmarks — and examines what it means for Oxford businesses competing in one of the country's most digitally concentrated markets.

The data on speed and money

The relationship between website load time and commercial performance has been studied extensively enough that the core findings are no longer seriously disputed.

Google and SOASTA's neural network research — which achieved 90% prediction accuracy using real-world browsing data — established the baseline figures that the rest of the industry now references. When page load time increases from one second to three seconds, the probability of a visitor leaving without interacting rises by 32%. Extend that to five seconds and the probability jumps by 90%. At ten seconds, the increase reaches 123% (Google/SOASTA Research, 2017).

These aren't theoretical projections. They're drawn from observed behaviour across millions of mobile sessions.

The Deloitte Digital study "Milliseconds Make Millions" — commissioned by Google and conducted in partnership with Fifty-five — went further by isolating speed as a variable and measuring its commercial impact across 37 brands in retail, travel, luxury, and lead generation sectors. Over a four-week period, analysing over 30 million user sessions, the team measured what happened when mobile load times improved by just 0.1 seconds across four speed metrics. The findings were consistent across every sector studied. Retail conversion rates increased by 8.4%, with average order values rising 9.2% — based on 20.5 million user sessions across 15 brands. Travel conversions rose by 10.1%. And luxury consumers proved the most speed-sensitive of all: a 0.1-second improvement produced a 40.1% increase in users progressing from product detail pages to adding items to their basket, drawn from 2.1 million sessions across 10 brands (Deloitte Digital / Google / Fifty-five, 2020).

The study is worth understanding precisely because of its methodology. The researchers didn't compare fast sites against slow ones — a comparison that would introduce countless confounding variables. They measured the same sites under naturally varying speed conditions and correlated marginal speed improvements with marginal changes in user behaviour. The 0.1-second improvement is important: it's small enough to be achievable through basic optimisation work, yet large enough to produce measurable commercial returns.

What the case studies show

Individual company data reinforces the aggregate research, and some of the most compelling cases come from well-known organisations that have published their results openly.

Vodafone's A/B test, documented by Google's web.dev team, is particularly instructive because of its controlled design. The company created two versions of a landing page that were visually and functionally identical — the only difference was that version A had been optimised for Core Web Vitals and loaded 31% faster on the Largest Contentful Paint metric. The optimised version produced 8% more sales, a 15% improvement in lead-to-visit rate, and an 11% improvement in cart-to-visit rate (web.dev, 2021).

The BBC's finding — that every additional second of load time costs 10% of users — led the organisation to implement automatic feature degradation under high traffic. When server load increases, lower-priority elements such as promotional modules are switched off to maintain speed (Creative Bloq / BBC, 2018). It's a revealing design choice: one of the world's largest media organisations decided that when speed and features conflict, speed wins.

Other documented cases follow the same pattern. Renault achieved a one-second improvement in Largest Contentful Paint and saw a 13% increase in conversions (web.dev, 2021). Swappie, the refurbished electronics marketplace, reduced LCP by 55%, CLS by 91%, and FID by 90%, producing a 42% increase in mobile revenue (web.dev, 2021). Agrofy improved LCP by 70% and saw load abandonment on product pages drop from 3.8% to 0.9% — a 76% reduction (web.dev, 2021). Tokopedia's performance optimisations led to a 35% increase in click-through rates and an 8% rise in conversions (web.dev, 2020).

None of these companies changed their products, pricing, or marketing. They made their existing websites load faster. The revenue followed.

Where the UK actually stands

Google measures website performance through Core Web Vitals — three metrics that assess how quickly a page loads its main content (Largest Contentful Paint), how quickly it responds to user interaction (Interaction to Next Paint), and how visually stable it is during loading (Cumulative Layout Shift). A page passes the assessment when at least 75% of real-user visits meet the "good" threshold for all three metrics.

According to the HTTP Archive and Chrome User Experience Report data from 2025, 57.1% of websites pass Core Web Vitals on desktop, but only 49.7% pass on mobile (DebugBear / HTTP Archive, 2025). That means more than half of all mobile websites are delivering a measurably poor experience by Google's own standards.

The 2025 Web Almanac — HTTP Archive's annual analysis of millions of websites — provides additional granularity. On home pages specifically, just 47% of desktop and 45% of mobile pages achieve good Core Web Vitals scores. Secondary pages perform better — 61% on desktop, 56% on mobile — likely because they benefit from cached assets and tend to be more templated (HTTP Archive Web Almanac, 2025).

Mobile pass rates have been improving — roughly 3 percentage points per year — but the gap between desktop and mobile remains persistent. Given that mobile now accounts for over 60% of web traffic and Google uses mobile-first indexing for search rankings, a website that passes on desktop but fails on mobile is effectively failing where it matters most.

In the UK specifically, websites load slightly faster than the global average. The median UK mobile page loads its main content in 1.8 seconds, compared to 1.9 seconds globally; on desktop, the UK sits at 1.6 seconds versus 1.7 globally (DebugBear / CrUX, 2025). These figures represent the 75th percentile of visitor experience — meaning one in four visits is slower. UK-specific CWV data shows that 29% of UK sites do not meet Core Web Vitals, while 70% achieve good Largest Contentful Paint scores.

WordPress — which powers approximately 43% of the world's websites and an even higher proportion of UK SME sites — presents its own performance picture. As of November 2025, 50% of WordPress sites achieve good Core Web Vitals scores on desktop, and 46% on mobile, based on analysis of over 3 million origins. That represents significant improvement from historical figures but still means that roughly half of all WordPress sites fail Google's performance assessment on mobile (HTTP Archive / WordPress.org, 2025).

What this costs in practice

The percentage-point statistics become more tangible when applied to actual business contexts.

The Deloitte study found that a 0.1-second improvement generated an 8.4% increase in retail conversions. For an Oxford retailer doing £500,000 in annual online revenue, an 8.4% improvement from a marginal speed gain represents £42,000. That's from a tenth of a second. The gains from moving a site from genuinely slow — four or five seconds — to genuinely fast are proportionally larger.

For service businesses, the impact works through enquiry volume rather than direct sales, but the mechanics are the same. If a solicitor or accountant's website sheds visitors during loading — and the Google/SOASTA data says it will, measurably, for every additional second — those are potential clients who never reach the contact page. Over twelve months, even a modest reduction in enquiry flow compounds into significant lost revenue. The precise figure depends on average client value, conversion rate, and traffic volume, but the direction is consistent across every study published in the last decade: slower sites generate fewer leads and fewer sales, and the relationship is not marginal.

Portent's analysis of over 100 million pageviews reinforces the point from the eCommerce side. Sites loading in one second convert at 2.5 times the rate of those loading in five seconds. For B2B sites, the gap widens to 3x (Portent, 2022). HubSpot's research adds that for every additional second of load time between zero and five seconds, conversion rates drop by an average of 4.42% — a compound penalty that means the difference between a two-second and a five-second site isn't 13%, but a fundamentally different conversion trajectory.

Why Oxford businesses face a higher bar

As we outlined in our analysis of Oxford's digital economy, the city's unusual concentration of knowledge-intensive businesses creates an environment where digital expectations are structurally higher than the national average.

Seventy-three percent of Oxford's employment is in knowledge-intensive sectors (Oxford City Council, Corporate Strategy 2024–28). These are sectors where decision-makers research online before making contact, where brand perception is shaped by digital experience, and where competitors include London firms actively marketing to Oxfordshire clients. A financial adviser in Oxford is competing not just against the firm down the road but against digitally sophisticated London practices whose websites load in under two seconds.

The tourism dimension reinforces the point. Oxford's visitor economy grew to £2.4 billion in 2024 despite an 11% drop in total visits — meaning the county is attracting fewer but higher-value visitors, particularly international ones (Experience Oxfordshire, 2024). International visitors research on mobile, often on variable-quality connections, and make booking decisions quickly. A tour operator or language school whose site doesn't render properly within three seconds on a mobile device is invisible to this audience.

For eCommerce businesses, the UK market's scale — £127.4 billion in 2024, the largest in Europe — creates both opportunity and pressure. Over 56% of UK eCommerce transactions happen on mobile (eMarketer/Statista). The Baymard Institute's meta-analysis puts average cart abandonment at 70.22%, with site speed among the leading contributory factors. In a market this mature, the businesses winning on performance are taking revenue directly from those that aren't.

What actually makes a website slow

Understanding what causes poor performance matters because many of the most common issues are straightforward to address — they're just not being addressed.

Unoptimised images remain the single largest contributor to page weight. The HTTP Archive reports that images make up nearly a third of a typical page's weight. Many business websites serve full-resolution images that are then scaled down by the browser — transferring far more data than the user ever sees. Converting to modern formats (WebP or AVIF), implementing responsive image sizing, and lazy-loading below-the-fold images can reduce page weight substantially with no visible quality loss.

Third-party scripts are the second major offender. Analytics tools, chat widgets, social media embeds, cookie consent banners, tracking pixels — each one adds HTTP requests and JavaScript execution time. The HTTP Archive found that 92% of websites load at least one third-party script, and the typical JavaScript payload grew 14% in the past year. Many of these scripts execute during page load, directly competing with the resources needed to render visible content.

Hosting and server response time determines the baseline from which everything else is measured. Google recommends a Time to First Byte (TTFB) of 0.8 seconds or less. Shared hosting — which most UK SME websites sit on — frequently exceeds this. The Hostingstep analysis found that half of all WordPress sites fail Core Web Vitals primarily due to poor server response times from shared hosting without edge caching. Upgrading to managed hosting with CDN support is often the single most impactful change a business can make.

Render-blocking CSS and JavaScript prevents the browser from displaying content until it has downloaded and processed these files. Deferring non-critical scripts, inlining critical CSS, and minimising the total amount of CSS and JavaScript are established optimisation techniques that many sites simply haven't implemented.

No caching strategy means repeat visitors download the same assets every time. Proper browser caching, server-side caching, and CDN implementation can reduce load times dramatically for returning users — who are, by definition, the most commercially valuable.

Core Web Vitals: the metrics that matter

Google's three Core Web Vitals provide a practical framework for measuring performance, and since they directly influence search rankings, they're worth understanding.

Largest Contentful Paint (LCP) measures how quickly the main content element — typically a hero image, heading, or text block — becomes visible. Google's "good" threshold is 2.5 seconds or less. This is the metric most directly tied to the "does this page feel fast?" question that determines whether users stay or leave.

Interaction to Next Paint (INP) replaced First Input Delay in 2024 and measures how quickly the page responds to user input — clicks, taps, keyboard actions. The "good" threshold is 200 milliseconds. A page that appears loaded but doesn't respond when you tap a button is arguably worse than one that's still visibly loading, because it violates user expectations.

Cumulative Layout Shift (CLS) measures visual stability — whether elements on the page move unexpectedly during loading. The "good" threshold is 0.1 or less. Layout shifts are the source of the maddening experience where you go to tap a link and the page jumps, causing you to hit something else instead.

All three metrics are measured using real-user data from the Chrome User Experience Report (CrUX), evaluated at the 75th percentile — meaning a site passes only if 75% of real visits meet the threshold. Lab tests using tools like Lighthouse are useful for diagnosis, but Google's ranking decisions are based on field data from actual users.

Where to start — and what to prioritise

Every optimisation guide on the internet will tell you to compress images and minify JavaScript. The harder question — the one businesses actually struggle with — is what to do first, and how to tell whether it's working.

Start with measurement, not fixes. Google's PageSpeed Insights gives you both lab data (a simulated test) and field data (real user experience from the Chrome User Experience Report). The field data is what matters for rankings. If your site has enough traffic to generate CrUX data, check your Core Web Vitals report in Google Search Console — it'll tell you exactly which pages are failing and why. If you don't have CrUX data, a Lighthouse audit is a reasonable starting point, but treat it as diagnostic rather than definitive.

Fix the infrastructure before the front end. The most common mistake is optimising images and scripts while sitting on shared hosting with a two-second server response time. If your Time to First Byte is above one second, front-end optimisation is fighting against a ceiling. A hosting upgrade or CDN deployment — Cloudflare offers a meaningful free tier — often delivers the largest single improvement for the least effort and cost.

Then address the biggest bottleneck, not all of them at once. PageSpeed Insights will identify your specific issues. For most SME websites, the priority order is: server response time, image optimisation, JavaScript reduction, then caching. Trying to fix everything simultaneously is how performance projects stall. Fix the top issue, measure the impact, then move to the next.

Treat performance as operational, not a project. Content updates, plugin changes, and third-party script additions quietly degrade performance over time. A site that scores well today can fail Core Web Vitals six months later without anyone touching the code. Even a monthly PageSpeed Insights check — free and takes two minutes — provides an early warning system. The businesses that maintain fast sites are the ones that monitor them.

The compound effect

Website speed isn't a standalone metric — it's a multiplier that affects every other digital investment a business makes.

Spend £5,000 on an SEO campaign to drive more traffic to a site that bounces 50% of mobile visitors, and you've wasted a significant portion of that investment. Run paid advertising to a landing page that loads in five seconds, and you're paying for clicks that never convert. Invest in beautiful design that takes seven seconds to render, and you've created an aesthetic that most of your audience will never see.

Conversely, improving load time from four seconds to two seconds doesn't just reduce bounce rates — it amplifies the return on every other pound spent on marketing, content, and design. The businesses seeing the highest ROI from their digital presence aren't necessarily the ones spending the most. They're the ones whose technical foundations ensure that spending isn't wasted.

This compounding works in both directions, and in Oxford's market it compounds faster. When 73% of local employment is knowledge-intensive, the audience is less forgiving of poor digital experience — they encounter high-performing websites all day in their professional lives, and they benchmark against them unconsciously. When international tourism visitors are spending more per head but arriving via mobile on variable connections, speed determines whether they find you at all. When London competitors are actively targeting Oxfordshire clients with sub-two-second sites, a local firm loading in four seconds isn't just slower — it's being outcompeted on a dimension the client may never consciously identify but will certainly act on.

The research assembled in this article spans a decade, dozens of companies, and hundreds of millions of user sessions. The conclusion is always the same. Speed is not a technical detail to be delegated and forgotten. It's a business metric with a direct, measurable, and compounding relationship to revenue. The only variable is how long a business waits before acting on it.

Oxford Web Services has been optimising website performance for Oxford businesses for over 20 years. If you'd like to know where your site stands, get in touch for a free performance review.

Sources

All statistics cited in this article are drawn from publicly available, verified sources:

• Google/SOASTA Research — Mobile page speed and bounce rate data (2017), via Think with Google
• Deloitte Digital / Google / Fifty-five — "Milliseconds Make Millions" report (2020), 30M+ user sessions across 37 brands
• web.dev — Vodafone Core Web Vitals case study (2021)
• web.dev / Creative Bloq — BBC load time and user abandonment data (2018)
• web.dev — Renault, Swappie, Agrofy, Tokopedia case studies (2020–2021)
• DebugBear / HTTP Archive — Core Web Vitals pass rates and CrUX data (2025)
• HTTP Archive — 2025 Web Almanac, Performance chapter (July 2025 data)
• HTTP Archive / WordPress.org — WordPress Core Web Vitals data (November 2025)
• DebugBear / Chrome User Experience Report — UK page load time data (2025)
• Hostingstep — WordPress Core Web Vitals statistics and hosting analysis (2025)
• Portent — "Site Speed is (Still) Impacting Your Conversion Rate" (2022), 100M+ pageviews
• HubSpot — Page load time and conversion rate correlation research
• Baymard Institute — Cart abandonment meta-analysis (2024–25)
• eMarketer / Statista — UK mobile commerce data (2024–25)
• Oxford City Council — Corporate Strategy 2024–28
• Experience Oxfordshire — Economic Impact Report 2024
• Google — Core Web Vitals thresholds and methodology (web.dev)