Google measures website speed with three numbers, and each one describes a different kind of slow. One is how long until the main thing on the screen appears. One is how quickly the page reacts when somebody taps it. One is whether the layout jumps about while it loads. A site can pass two and fail the third, and the cause is usually something specific rather than a general need to make things faster.
How long until the page looks loaded
This measures when the largest visible thing, usually a heading or a hero image, finishes drawing. Under two and a half seconds is the target.
The usual cause of failure is one enormous image. A photograph straight from a camera, three thousand pixels wide, shown at eight hundred pixels, carries more than ten times the pixels the screen can use.
Second most common is a stack of third party scripts loading before anything visible does. Analytics, chat widgets, tracking pixels and font loaders all queue ahead of the content when they are put in the wrong place.
Hosting distance matters here too. Every request to a server on another continent carries the round trip with it.
How quickly the page answers
This measures the delay between somebody tapping something and the page responding. Under two tenths of a second is the target.
Failure here is almost always too much JavaScript. The browser is busy running code, so the tap waits its turn.
On a site assembled from many plugins this is cumulative: each one adds a little, none of it is anybody's fault individually, and the total is a page that feels sticky on a mid range phone.
Whether the page moves while you read it
This measures layout shifting. It is the thing that makes you tap the wrong button because an advertisement appeared and pushed everything down half a second after you aimed.
The fix is almost mechanical: give every image and every embed a stated width and height, so the browser reserves the space before the content arrives.
The other common cause is a web font swapping in and changing every line length on the page at once.
What to do about it, in order
Measure on a phone, on a normal connection. A desktop on office wifi tells you almost nothing about what your customers experience.
Fix the images first. It is the largest gain for the least work on nearly every site: correct dimensions, a modern format, and lazy loading for anything below the fold.
Then count the third party scripts and remove the ones nobody looks at. Most sites are carrying at least one tracking tool nobody has opened in a year.
Then set dimensions on everything that loads late.
Only then start on the harder work. Most sites never need to, because the first three steps have already got them there.
Questions people ask
- Does site speed affect Google rankings?
- Yes, though less than people assume. It is one signal among many and it rarely outweighs relevance. Where it matters much more is in whether people stay: a slow page loses visitors before they read anything, and that costs more than the ranking difference does.
- What is a good score?
- Scores are a summary, and chasing a perfect one is usually wasted effort. What matters is passing the three measurements on a real phone. A site that passes all three with a score in the eighties is in better shape than a site with a higher score that fails one of them.
We do this
If you would rather have it done than do it, this is the part of what we offer that covers it.