Your Site Is Fast on Your Laptop and Slow Where It Counts

Ask a business owner whether their site is fast and almost all of them say yes. They are not wrong about what they have seen. They are looking at a site they have visited a hundred times, on a desktop machine, on the office connection, with every image already sitting in their browser cache. It opens instantly because for them it practically already has.

The customer is on a phone that is three years old, on cell data, standing in a hallway, having tapped a result they have never visited before. Nothing is cached. That is the load that decides whether you get the inquiry, and it is the one nobody in the building ever experiences.

What slow actually costs

The damage is not evenly spread. People who already know your name will wait — they came looking for you specifically and a few seconds will not send them elsewhere. The people who leave are the ones who found you in a list of options, which is to say the new customers, which is to say the entire point of the marketing.

There is a second cost that shows up nowhere in your analytics. If the page never finished loading, the visit often never registered, so a slow site under-reports exactly the traffic it is losing. The graph looks fine because the people it should be worried about never made it into it.

Measure it the way a stranger would

Two checks, five minutes, and you will know more than most audits will tell you.

1. Run your busiest page through Google's PageSpeed Insights and read the mobile tab, not the desktop one. Ignore the overall score for a moment and look at Largest Contentful Paint — how long until the main thing on the screen is actually there. Under two and a half seconds is healthy; past four, you have a problem worth money.
2. Open the site on your own phone, with wi-fi switched off, on a page you have not visited recently. Count out loud. What you are timing is not the whole page but the moment it stops being blank and the moment it stops moving around under your thumb.

If the field data section of that report has numbers in it, trust those over the lab score. It is drawn from real visits on real devices, which is the only measurement that has ever mattered.

The four things that are usually to blame

On small business sites the causes repeat with remarkable consistency, and they are roughly in this order of impact.

• Images shipped at their original size. A photograph straight off a phone camera can be several megabytes and four thousand pixels wide, displayed in a box six hundred pixels across. Resizing and converting the handful of images on your busiest pages is usually the single largest win available, and it takes an afternoon.
• Third-party scripts. Chat widgets, tracking tags, review carousels, booking embeds, a font service, a popup tool. Each one is code loaded from somebody else's server, and your page waits on all of them. Most sites are carrying at least two nobody remembers adding.
• A page that rearranges itself while loading. Text appears, then an image arrives above it and shoves everything down just as a thumb comes in to tap. This is measured as Cumulative Layout Shift, and it is experienced as the site being broken rather than slow.
• Hosting and plugin weight. A stack of plugins each doing a little work on every request adds up, and cheap shared hosting adds a delay before the first byte that no amount of front-end tuning can recover.

Fix in this order

Start with images on the three pages that receive the most traffic, because the effort is low and the effect is immediate. Then audit the third-party scripts and remove anything that is not earning its keep — the test is whether you could name what it does and what it produced last month. Then set explicit dimensions on images so the layout stops jumping. Only after that is it worth arguing about hosting.

Do not chase a perfect score. The number is a diagnostic, not a goal, and the last few points typically cost more than they return. What you are after is a page that shows its main content quickly and then holds still.

Speed is not a technical metric. It is the first impression, delivered before a single word of your copy has been read.

Speed alone will not save a page that has nothing to say

A fast page that fails to answer the question someone arrived with is a fast failure. Plenty of sites load in under two seconds and still convert badly, because the thing a visitor needed — the price range, the service area, the proof, an obvious way to get in touch — was not on the screen. Speed removes an obstacle; it does not supply an argument. That distinction is the whole of why your website gets visitors but no leads.

The right order is to make the page work and then make it fast, which is why performance is part of the build rather than a later cleanup in the websites we develop. If you want a plain read on where your current site is losing people, before the fix list and before the invoice, ask for a proposal.