How we measure
Where every number on your result came from
We did not invent a scoring system. The figures for your site come from Google's own tool, and our own figures are measured on the same tool, by hand, with the date stamped on them. Here is the whole method, including the parts that are less flattering to us.
Two kinds of number, and we always say which
Google can answer the question two ways. One is much stronger than the other, and which one you get depends on how busy your site is rather than on anything you have done.
Google quietly records how quickly pages load for actual Chrome users, and publishes it as a 28 day summary. If your site gets enough traffic to qualify, that is what we show you.
- Real people, real phones, real broadband and 4G
- Averaged over four weeks, so one bad day cannot skew it
- The same data Google uses when it decides how to rank you
- Needs a reasonable amount of traffic before it exists at all
If your site is quieter than that threshold, and plenty of good small business sites are, there is no visitor data to show. So we fall back to loading your page once on a simulated phone and measuring that.
- Available for any site, however new or quiet
- Good at finding the cause of a problem
- One machine, one connection, one moment, so it moves between runs
- Not proof of what your customers actually get
Every result is badged with which of the two it is, at the top of the page, before any of the numbers. Having no visitor data is not a mark against your site. It usually just means you are a smaller business than the threshold Google set. Our own homepage is below it too, so our column is a single test run as well.
How our own figures are produced
The two questions a sceptic asks about the other column are when it was taken and whose site it is. Both are answered here rather than left to be assumed.
Our own homepage, on a stated date
The figure you are compared against is coullweb.com's own homepage, put through Google PageSpeed Insights exactly as your site is. It is one page, ours, and we name it — you can run the same test on the same address yourself and check.
Google holds no real visitor data for us either. We are below its traffic threshold, the same as most small business sites, so our column is a single test run and yours may well be too.
Our own homepage, measured 5 August 2026
Refreshed by hand, not on a timer
There is no scheduled job behind our number. We measure our homepage, record what it said and the day it said it, and refresh it when we change the site. The date is stamped on the figure everywhere it appears, so you can see for yourself how current it is.
We do not run ours live alongside yours for two reasons: it would double how long you sit waiting, and one unlucky run on our side would make the comparison meaningless in the other direction.
The same test, both ways
Your site and ours go through Google PageSpeed Insights on the same settings: mobile first, throttled connection, the homepage rather than a page we chose. If you paste a deep link, we test the homepage of that address instead and say so on screen before you press go.
Where Google holds real visitor data for a site, that is what is used, for you and for us alike.
Results are held for a day
We keep each result for 24 hours against the address that was tested. If you or somebody you shared the link with opens it again in that window, you get the stored result straight away rather than a fresh wait.
It is always labelled with the time it was taken, and you can force a fresh run whenever you like.
What this tool will not do
It is not a formal audit
It measures how quickly one page loads and nothing else. It does not check whether your site is secure, whether it can be found, whether it works on every device, or whether it is any good at bringing you work. Do not take a good result here as a clean bill of health.
It is for your own site
Test a site you own or look after. We are not going to produce a report about a named competitor for you, we do not publish tables of other people's scores, and nothing you test here appears anywhere public.
It tests the homepage, not the whole site
One address gets you one page: the homepage. Paste a deep link and we normalise it to the homepage of that site rather than testing the page you pasted, and we tell you that on screen before the test runs. A slow product page will not show up here.
The number is not a promise
Speed changes. A plugin update, a new photo on the homepage or a busy afternoon can all move it. A result from today is a snapshot of today.
There is a limit on how many you can run
Five an hour from the same connection. The tool is free and the tests cost us something to run, so the cap is what keeps it free and available. If you look after a lot of sites and need more, email support@coullweb.com and ask.
Ready to see your own numbers?
About a minute, and no email needed.