Marketing website performance: what fast pages are actually worth
By Nayan Kyada · · 6 min read
Part of Next.js Performance
Marketing website performance is one of those topics that sounds technical until you translate it into revenue numbers — then it becomes urgent. If your site takes four seconds to show its main headline on mobile, you are paying for ad clicks that bounce before anyone reads a word. This post explains what the key metrics mean without jargon, what the research says about conversion impact, and the specific questions worth asking any agency or developer before you sign.
Why load time is a business metric, not a developer metric
Google published data years ago showing a 32% higher bounce probability when page load time goes from one second to three seconds. More recent industry benchmarks put the conversion penalty at roughly 4–8% per additional second on e-commerce and lead-gen pages. Those figures are averages across all industries — SaaS landing pages, where visitors arrive from paid search with high intent, tend to be more sensitive, not less.
The compounding effect is the part most site owners miss. A slow page hurts you three times: the visitor bounces before converting, your Quality Score on Google Ads drops (raising your cost per click), and Google Search applies a ranking signal that gradually pushes slower pages down over organic results. You pay more to acquire less traffic and then convert a smaller fraction of it.
Core Web Vitals translated into plain language
Google formally measures site experience through three signals called Core Web Vitals. They are reported in Google Search Console for free, and they feed into ranking. Here is what each one means without the acronym soup:
How fast does the main content appear? This is Largest Contentful Paint (LCP). For most marketing pages, the LCP element is the hero image or headline. Google's threshold for a "good" score is under 2.5 seconds on a real mobile connection. A slow LCP usually means large uncompressed images, no caching, or a slow hosting provider.
Does the page jump around while loading? This is Cumulative Layout Shift (CLS). If your headline moves down when a banner loads, or the CTA button shifts just as someone tries to click it, that is CLS. It is irritating, it destroys trust, and it is surprisingly common on sites built with off-the-shelf page builders. Google wants a score below 0.1.
How quickly does the page respond when someone interacts? This is Interaction to Next Paint (INP). It replaced the older First Input Delay metric in 2024. In practice, INP measures how long it takes for the page to visually respond after a tap or click. Pages that load a lot of JavaScript — analytics tags, chat widgets, third-party embeds — often score poorly here even when they look fast.
All three metrics are measured on real visitor data from Chrome users, not a lab test. That matters because a developer running PageSpeed Insights on a fast laptop in a city with fibre broadband will see much better numbers than a first-time visitor on a budget Android phone in a patchy 4G area.
What fast pages are actually worth
The ROI varies by traffic source and page goal, but some benchmarks are grounded enough to use in a business case:
| Scenario | Reported impact |
|---|---|
| 1-second improvement in mobile load time | +2–5% conversion rate on lead forms (Portent, 2024) |
| LCP under 2.5 s vs over 4 s | ~15% higher form completion rate (Google internal data) |
| CLS score reduced from 0.25 to under 0.1 | Reduced accidental click bounces, lower exit rate |
| INP improvement on SaaS trial page | Measurable reduction in abandoned sign-up flows |
These are averages. Your site may do better or worse depending on audience and page design. But if your monthly ad spend on the pages in question is £5,000 or more, even a 2% conversion improvement pays for a serious performance audit inside a month.
What to demand from any agency or developer
Before you hire someone to build or improve a marketing site, ask these questions. Vague answers are a signal.
Can you show me Core Web Vitals scores from a recent site you built — field data, not lab? Field data appears in Search Console and CrUX (Chrome User Experience Report). Lab scores from PageSpeed are easier to manipulate with a few quick optimisations that do not survive real traffic patterns. A developer who cannot produce field data from a live site either has not shipped enough sites or has not monitored them.
How are images served and sized? The correct answer involves modern formats (WebP or AVIF), responsive srcset attributes so mobile devices do not download desktop-sized files, and explicit width and height on images to prevent layout shift. If the answer is "we use a CDN" without more detail, push harder.
What third-party scripts are on the page by default, and how are they loaded? Analytics, chat, heatmaps, and ad pixels all have a cost. Each one adds to INP risk. A performance-aware developer can defer non-essential scripts so they load after the page is interactive, rather than blocking it. Ask for a list of what ships by default and what the loading strategy is.
How is caching handled, and what is the expected Time to First Byte? For a standard marketing page with no personalisation, there is no good reason for TTFB to exceed 200 ms. If the answer is "it depends on the server" without a caching or edge delivery strategy, the site will likely underperform on mobile real-world tests.
What happens when content editors add a large image? If the build has no guard against an editor uploading a 6 MB JPEG that goes straight to production without compression, your LCP score will degrade the moment the team is working independently. A well-configured CMS pipeline (Sanity's image pipeline, for instance, or Contentful's image API) handles this automatically. Verify it is actually configured, not just available.
The handoff question most buyers forget
Performance can be built into a site at launch and then degraded within weeks if the CMS setup, image handling, and third-party script policy are not part of the handoff documentation. When a developer or agency hands over a site, ask for the Core Web Vitals baseline scores from Search Console, a written explanation of how images are optimised, and confirmation that the analytics and marketing tag setup has been reviewed for performance impact.
A fast marketing site is not a one-time delivery. It is a configuration that needs to survive real editors and real content. The questions above will help you distinguish developers who understand that from those who optimise for the demo.