Contents
01Measure it the way your visitors experience it
Test on a mid range phone over a mobile connection, not on the machine you built the site on. The gap between those two is where most arguments about speed come from: the site genuinely is fast for the person who made it.
Google publishes the thresholds it uses, and the main one is simple. The largest piece of content should appear within 2.5 seconds. The page should respond to the first tap quickly, and it should not shift under a thumb while it finishes loading.
02Where the seconds actually go
In the sites we audit, the order is remarkably consistent.
- Images sent at full size and in old formats. One hero image is often heavier than the entire rest of the page.
- Fonts that block text from appearing, especially several weights of several families.
- Third party scripts: chat widgets, heat maps, tag managers loading more tags than anyone remembers adding.
- A page that renders in the browser when it could have arrived as finished HTML.
Open your tag manager and count the tags nobody can explain. That audit alone frequently returns half a second.
03The fixes, in the order they pay
Start with images: correct sizes, modern formats, and dimensions set in the markup so nothing jumps. Then fonts: fewer weights, loaded so text appears immediately. Then the scripts: remove what nobody reads, defer what is not needed for the first screen, and move analytics behind consent where it belongs anyway.
Only after that does it make sense to discuss rebuilding. Most sites reach acceptable numbers without one, and the ones that do not usually have a deeper problem: a page assembled in the browser from several slow sources.
04What it is worth in money
Work it out with your own numbers rather than an industry statistic. Take your monthly ad spend, your current enquiry rate and your share of mobile visitors. If a tenth of mobile visitors leave before the page appears, you are paying for that tenth every month and will keep paying.
Our audit of accessibility and performance is $2,600 and produces a prioritised list with the measured effect of each item. Several clients have implemented the first half themselves, which is a fine outcome: the list is yours.
Questions
Does speed actually affect rankings?
It is one signal among many, and it rarely moves a page from the second page to the first on its own. It affects revenue more directly than rankings, because people leave slow pages and you already paid for them to arrive.
Our site scores well in the lab but feels slow. Why?
Lab tools run on a simulated connection with nothing else happening. Field data, which is what Google uses, comes from real visitors on real phones with other tabs open. Trust the field numbers when they disagree.
Is a rebuild the only way to fix a slow site?
No, and we will tell you when it is not. Images, fonts and scripts usually account for most of the loss. A rebuild is warranted when the page is assembled in the browser from several slow sources, because no amount of tuning removes that.
Where we can help