Website Speed and Core Web Vitals: Why a Slow Site Loses Customers
• Wednesday, 26th August 2026 • Web development •
The customer does not wait; they go back to the search results
The behaviour of a user who lands on a local business’s website is brutally simple. They click a search result, look at a white screen, count to three in their head, and go back to Google to click the next result. From the business owner’s perspective there is no warning signal here: no phone rang, no email arrived, nothing appeared that could be noticed. The enquiry simply never came into being.
User behaviour research has shown the same thing for years: every extra second of loading raises the abandonment rate, and the biggest jump comes between the second and fourth second. On mobile devices, which generate most of the traffic for local services, the effect is even stronger, because the connection is often worse and patience is shorter.
Core Web Vitals without the jargon
For several years Google has measured three indicators that describe the real user experience. Each answers a different question.
LCP: when the content is visible
Largest Contentful Paint measures the time after which the largest element in the visible part of the page appears on screen: usually a header image or a large block of text. A value below two and a half seconds is considered good, and above four seconds needs improvement. It is the indicator users feel most strongly, because it answers the question “is there anything here at all”.
INP: whether the page responds
Interaction to Next Paint replaced the earlier first input delay metric and measures how quickly the page responds to the user’s actions throughout the whole visit. Clicking the menu, expanding a tab, submitting a form: if the reaction comes after more than half a second, the page feels frozen. A good result is below two hundred milliseconds.
CLS: whether the content slips from under your finger
Cumulative Layout Shift describes how much elements jump around during loading. The classic situation: the user starts reading, a banner loads in, the text slides down, and their finger hits the wrong button. The score should be lower than one tenth.
What most often slows down small business websites
In audit practice the same few causes keep recurring, and it is rarely the “weak server” that suspicion usually falls on first.
- Unscaled images: a photograph straight from the camera, five thousand pixels wide and weighing six megabytes, displayed in a window eight hundred pixels wide. This is the most common single problem.
- No modern formats: the same image saved as WebP usually weighs about half as much as a JPEG at identical quality visible to the eye.
- Too many plugins: in a typical business website, half of the installed extensions do nothing visible, and each adds its own scripts and styles.
- External scripts: chat widgets, maps, social media feeds and consent systems. Each of them is a separate connection to another server.
- Web fonts: three font families in five variants can add several hundred kilobytes and delay the appearance of text.
- No caching: a returning user downloads the whole page again instead of using stored files.
Where to start fixing it
Optimisation gives the best return when it starts with the heaviest things, not the most technically impressive ones.
- Measure the baseline: Google’s tools show lab results and data from real users. The latter matter more.
- Deal with images: scale them to the actual display dimensions, convert to WebP, add lazy loading below the first screen, and always specify dimensions so the browser reserves space.
- Review plugins and scripts: switch off what you do not use. A chat widget that nobody uses costs real loading time on every subpage.
- Limit fonts: two families in two weights are enough in most projects.
- Enable compression and caching on the server side. This is usually a matter of configuration, not rebuilding the site.
- Check the hosting: only at the end. If after the steps above the server response time still exceeds half a second, the infrastructure really is the problem.
Does speed really affect Google rankings?
Yes, but not in the way it is usually presented. Core Web Vitals are one of many ranking signals and will not outweigh the relevance of the content. A slow page that answers the query precisely will beat a fast and empty one. The difference appears between sites of similar content quality, and that is exactly where local results are decided, where several businesses from the same town offer the same service.
The indirect effect, however, is far more important. A faster page means a lower abandonment rate, a longer visit and more completed forms. These signals translate into business results regardless of how Google treats them in rankings.
How much does it cost and when does it pay back?
Optimising an existing website usually costs somewhere from a few to a dozen or so hours of work, depending on the starting condition and the technology. For a site generating a dozen or so enquiries a month, recovering even two of them pays back that cost in the first quarter. A rebuild from scratch is justified only when the site is several years old, relies on an abandoned theme and every fix requires working around the template’s limitations.
If you suspect that your website is putting off some customers before they have had time to read anything, we can measure it and show you specifically what slows it down and how much can be recovered without a rebuild. Write to us, and we will prepare a speed audit together with a list of fixes ranked by their real impact on results.
