It is natural to wonder how your website stacks up against the competitor down the road. The usual approach is to paste both addresses into PageSpeed Insights and compare the two numbers. It takes a minute, and it can easily tell you the wrong thing.

The short answer: compare the same page type, on the same device, more than once, and treat small gaps as no gap at all. Use the result as context for your own site's data, not as a verdict on who is "winning".

Why one score misleads

A PageSpeed score comes from a lab test: Google loads the page once, on a simulated phone or computer, and scores it from 0 to 100. We explain the difference between that and what your visitors experience in why your PageSpeed score differs from what your visitors experience. For a comparison, three problems matter most:

Run-to-run noise. The same page can score several points differently on consecutive runs. A gap of three or four points between two sites is well inside that noise.
Different pages. Your homepage may be a simple page while a competitor's is a busy shop front. Comparing unlike pages tells you about the pages, not the sites.
Different devices. A mobile result and a desktop result are not comparable, and mobile scores are usually lower.

A fairer way to compare

Pick like with like. Compare your homepage with the competitor's homepage. If the competitor only has one page that matters (a profile on a platform, for example), compare with that page, and remember that it is not a like-for-like test.
Use the same device. Run mobile against mobile and desktop against desktop.
Run each page a few times. Note the range, not one lucky or unlucky score.
Look past the headline number. The individual timings, such as how long the main content takes to appear, say more than the overall score.
Ignore small gaps. If the scores are within a handful of points, call it level.

You can do all of this by hand at PageSpeed Insights. It is slower than it sounds, and you have to remember to repeat it.

Real-visitor data, when Google has it

For some sites, PageSpeed Insights also shows Google's real-visitor figures from the Chrome UX Report. When both sites have them, these are the fairer comparison, because they reflect real people rather than a single simulated visit. Compare the category Google gives each metric (fast, average or slow) rather than the exact numbers.

Many smaller sites have no such data. When either side is missing it, do not guess: compare only the lab results and say so.

What a comparison cannot tell you

Traffic. Speed data says nothing about how many people visit a site. Google's real-visitor data only shows that it has enough measurements, not how many visits there are.
Rankings. A faster page is not guaranteed to rank higher, and a slower one is not guaranteed to rank lower. Content and links matter a great deal.
Which site is "better". A comparison describes two pages on one day. It does not describe a business.

How the structure of a site compares

Speed is only one way to look at a competitor. Another is how the site is put together: one long page with sections, or many separate pages? A site's sitemap and home page show how many pages it lists, how many sections they fall into and what its main navigation says.

These are counts, not scores. More pages or a deeper structure is not better or worse in itself. It is useful as context, for example when you are deciding whether a one-page site suits what you are trying to do.

How we do it

If you add up to three competitor sites, or single pages, to Web Vitals Collector, the weekly report adds a Competitors tab. It compares your homepage with each one on mobile and desktop using Google's PageSpeed results, treats small gaps as "similar", compares Google's real-visitor figures only when both sides have them, and shows how each site is organised. It is a comparison with the sites you choose, not a ranking, and your own visitors' measurements still decide whether your site passes.

To keep that fair we only read public information: for speed, Google fetches the page; for structure, our crawler reads the competitor's robots.txt, sitemap and home page and respects what robots.txt says. See about our website crawler for exactly what it reads and how to block it.

You can see what a weekly report looks like in our sample report.

PageSpeed and Lighthouse results are automated estimates and are not a guarantee of search ranking. Competitor results depend on third-party websites we do not control.