Site speed gets treated as a technical vanity metric. It is not. It is the number of people who leave before they have seen anything.
Nobody has ever left a website because the performance score was 62. They leave because they tapped a link, looked at a blank white screen for four seconds, and went back to the search results where three of your competitors were waiting.
That is what speed is. Not a score. A queue at the door, and the people at the back of it leave.
The three things Google actually grades
Google publishes three measurements it uses, and they are worth knowing because each one maps to something a visitor genuinely experiences.
- Largest Contentful Paint. How long until the main thing on the screen appears. This is the one people feel. Under two and a half seconds is the target.
- Cumulative Layout Shift. Whether the page jumps around while it loads. This is the reason you tap the wrong button and land on the wrong page. It is the most irritating of the three and often the easiest to fix.
- Interaction to Next Paint. How long the page takes to respond when you tap something. A page that looks ready and ignores your first two taps feels broken even when it is not.
What usually makes a small business site slow
In our experience it is nearly always the same four, in roughly this order.
Images. A photo straight off a phone can be four or five megabytes. Put six of them on a home page and you have built something that takes half a minute to load on cell data. Resizing them to the size they are actually displayed at, and serving them in a modern format, routinely cuts a page weight by three quarters. It is the single biggest win available and it changes nothing about how the site looks.
Fonts. Custom fonts are fine. Five families in nine weights, loaded from somebody else’s server, are not. And a font that arrives late makes the whole page reflow, which is the layout shift problem above.
Scripts. Chat widgets, popups, heat maps, three analytics tools, two ad pixels, a review carousel. Each one seemed small. Together they are usually the largest thing on the page, and half of them are for tools nobody has opened in a year.
Hosting. The cheapest shared hosting is cheap because your site shares a machine with hundreds of others. It shows up as a slow first response before a single image has downloaded.
The fix nobody wants to hear
Most of the real gains come from removing things rather than optimizing them. The fastest slider is no slider. The fastest popup is the one you deleted. The fastest tracking script is the one for the tool you stopped using in March.
Go through what is loading on your home page and ask, for each item, what it earned last month. Most of them will have earned nothing, and they were costing you the people at the back of the queue.
Mobile is the real test
Nearly everyone testing their own site is doing it on a fast connection on a desktop, sitting in the building where the site is hosted. That is the best case and it is not the case that matters.
Test it on a phone, on cell data, away from your wi-fi, on a page that is not the home page. If that feels fine, you are in good shape. If it does not, that is what a real customer in Lexington has been getting.
How to measure it yourself
Google’s PageSpeed Insights is free, takes a URL, and gives you both the lab test and, if you have enough traffic, real data from actual visitors. Use the mobile tab. Ignore the overall number and read the three measurements above and the list of what is biggest.
One caution: a single run is not a measurement. Scores on the same unchanged page can move by ten points or more between runs. Run it three times and take the middle one, and compare like with like, same page and same device type, or you will chase noise for a week.
What good looks like
For a small business site: the main content visible in under two and a half seconds on a phone, no visible jumping as it loads, a total page weight under about a megabyte and a half, and fewer than a couple of dozen requests before the page is usable.
That is achievable on almost any budget. It is mostly a question of somebody being responsible for it after launch, which is the part that tends to go missing. If you would rather it was watched rather than remembered, that is part of what our growth program and our website work are for.