PageSpeed Results in Plain English

Runs Google's PageSpeed Insights test on a page and explains the report in plain English: a short list of fixes, the ones Google expects to save the most time first. For founders and marketers who get a PageSpeed report and do not know which of its dozens of lines matter. The measurements and estimates are Google's; Reavlo only explains them.

How to use it

Enter the address of one page and choose whether to test it as a phone or a desktop computer.

1. Type or paste the page address, choose Mobile or Desktop and press Run. The test takes 10 to 30 seconds.
2. Read the numbers Google measured: when the main content appears, how long the page ignores taps and how much it jumps.
3. Start with the critical and warning findings. They are ordered so the fix Google expects to save the most time comes first.
4. Make one change at a time and run the test again to see whether it helped.

How Reavlo tests this

This tool runs on our server and uses another company's service: our server sends the address you enter to the Google PageSpeed Insights API, which loads the page in a test browser and returns a Lighthouse report. The API key stays on our server. Reavlo stores nothing from the run, but Google receives the address; see the privacy page.

What is Google's and what is ours. Every measurement, every estimate of time or size saved and the performance score come from Google's report and are shown as Google's. Reavlo computes no score of its own. What Reavlo adds is the reading: each Lighthouse audit that Google scores below 0.9, the line where Lighthouse itself marks a pass, becomes a finding with a plain-English fix, and the findings are ordered by severity and then by the time Google estimates the fix would save.

One test is one sample. The result depends on Google's test machine, its network and the moment of the test, so two runs of the same page differ. Treat a small difference as noise and look at the findings that keep coming back. Mobile is the stricter test; it simulates a mid-range phone on a slower connection.

Lab test, not real visitors. PageSpeed Insights also publishes data from real Chrome users for some sites, and this tool does not show it. A page can pass the lab test and still be slow for people in other places, or the other way around.

What the rules cover. Fifteen rules cover the main metrics (largest content, first content, blocking time, layout shift, server response) and the common causes: render-blocking files, images, unused or unminified code, missing compression, short caching, page size, too many elements and redirects. Lighthouse checks more than this; a rule is added only when its fix can be explained in one sentence.

When Google cannot run the test. Google limits how many tests our server can run each day. When the limit is reached, when Google cannot load the page or when its service is down, the tool says so and links to the same test on PageSpeed Insights, where you can run it yourself.

Questions