Why a Slow Website Costs You Customers
Most local sites take five or six seconds to open on a phone. Half the visitors are gone before the page appears, and they never show up in your analytics.
Speed is the least interesting thing about a website and one of the most expensive to get wrong. Nobody ever chose a plumber because the site loaded quickly. Plenty of people have chosen a different plumber because it did not.
The reason it gets ignored is that the damage is invisible. A visitor who leaves before the page renders never appears in your analytics as a lost customer, because they never became a visitor at all. You just see fewer calls than you expected and no obvious reason why.
What the delay actually costs
The pattern is well established across the industry and matches what we see locally: drop-off climbs sharply with each additional second, and it is worse on mobile than desktop.
| Load time | What it feels like | What happens |
|---|---|---|
| Under 1s | Instant | Nobody leaves. The page is there before they look up. |
| 1–2.5s | Fast | Normal behaviour. This is the target. |
| 3–4s | A noticeable wait | A meaningful share leave, more if they arrived from a search result. |
| 5s+ | Broken | Most visitors are gone. Many assume the business is closed or the site is dead. |
The last row is where a surprising number of local sites sit. It matters most for exactly the traffic you want — someone in a hurry, on mobile data, with three other search results still open behind your tab.
The four usual causes
1. Images straight off a camera
This is the biggest one by a distance. A photo from a modern phone is four to eight megabytes. Uploaded without compression and displayed in a small box on the page, the browser still downloads the whole thing. Ten of those on a homepage is fifty megabytes for a page that should weigh under two.
Compressed and served in a modern format, the same photos are typically a twentieth of the size with no visible difference.
2. Page builders doing too much
Drag-and-drop builders load the machinery for every feature they offer, whether or not your page uses any of it. A five-page brochure site ends up shipping the code for sliders, popups, animations and a shop it does not have.
3. Third-party scripts
Chat widgets, tracking pixels, review carousels, booking embeds, consent banners, analytics. Each is a request to somebody else's server, and each can block your page while it waits. Four or five of them is common, and they compound.
4. Fonts
Loading four weights of three font families is roughly a megabyte before a word appears, and badly configured fonts leave text invisible while they download. Two families, two or three weights, with a sensible fallback, is plenty.
Test your own site in two minutes
- Open PageSpeed Insights and enter your homepage. Read the mobile score, not the desktop one — desktop flatters everything.
- Look at Largest Contentful Paint specifically. Under 2.5 seconds is good, over 4 is a problem.
- Check the opportunities list. Image sizing is usually at the top, and usually the easiest fix.
- Then do the real test: open your site on your own phone, on mobile data with wifi off, somewhere with a weak signal.
What to fix, in order
- Compress every image. Usually the single biggest win and can be done in an afternoon without touching the design.
- Remove the scripts you do not use. The abandoned chat widget, the old analytics tag, the review carousel nobody looks at.
- Cut the fonts back. Two families, a couple of weights, with a real fallback stack.
- Load the page in a sensible order. Text and the phone number first, decorative images and widgets last.
- Then consider the platform. If the weight is structural, no amount of tuning fixes it, and a rebuild is the cheaper answer.
Where this matters most
Speed is not equally urgent for everyone. It is close to critical if you sell to people in a hurry or in poor signal.
- Plumbers and HVAC companies, because the customer is in a basement with one bar and a leak.
- Restaurants, because the customer is on a crowded street deciding between you and next door.
- Any business in Mount Pleasant during move-in week, when campus networks are saturated.
- Any business in Midland, where the audience compares you daily to fast corporate software.
If your site takes more than three seconds on a phone, that is worth an afternoon before you spend anything on advertising. Sending paid traffic to a slow page is paying to watch people leave.
Every site we build is targeted at a sub-two-second paint on a mid-range phone, and we check it on real devices before launch. If you want a second opinion on the site you already have, we will test it and tell you what we find, including when the answer is that it is fine.
Frequently asked
How fast should a small business website be?
The practical target is meaningful content on screen in under two seconds on a mid-range phone on mobile data, and under 2.5 seconds for Largest Contentful Paint, which is the metric Google measures. Most local business sites we test are between four and eight seconds, so the bar is lower than it sounds.
Does page speed actually affect Google rankings?
Yes, but indirectly and less than people assume. Core Web Vitals are a confirmed ranking signal, though a fairly light one. The larger effect is behavioural: slow pages make people bounce back to the results and click a competitor, and that pattern does real damage over time.
What usually makes a small business site slow?
Almost always one of four things: uncompressed images straight off a phone or camera, a page builder loading scripts for features the page does not use, third-party tags like chat widgets and tracking pixels, and unoptimised fonts. Images alone account for the majority of what we see.
Can a slow site be fixed, or does it need rebuilding?
Often it can be fixed. Compressing images and removing unused third-party scripts frequently takes several seconds off without touching the design. What cannot be fixed cheaply is a site built on a heavy page builder with thirty plugins, where the weight is structural rather than incidental.