How to test your website's speed, and which number actually matters

Updated November 14, 2026 · 5 min read
Short answer

Run your key pages through Google's free PageSpeed Insights on the mobile setting and read the top section first: the Core Web Vitals assessment from real visitors. That pass or fail result is the number that matters. The 0 to 100 performance score below it is a lab simulation, useful for finding problems but not the grade Google uses.

Two kinds of speed data, and why they disagree

Every speed tool reports one of two things. Field data comes from real visits to your site over the previous 28 days, collected from Chrome users who share anonymous usage statistics. Lab data is a single simulated visit, run by a tool called Lighthouse on a deliberately slowed connection and processor.

They often disagree, and that is normal. The lab uses one device, one network and an empty cache. Real patients use every kind of phone, sometimes have parts of your site cached from an earlier visit, and tap buttons at moments a lab cannot predict.

Google's guidance is clear: when you have both, use field data to decide what to fix. Use lab data to find out why, and to check a fix before the real user data catches up.

PageSpeed Insights, step by step

  1. Go to Google's PageSpeed Insights and paste the full address of one page, starting with your homepage.
  2. Leave it on the Mobile tab. Many patients search on phones, and mobile is the harder test.
  3. Read the top section first. This is the real user data, and it shows whether the page passes the Core Web Vitals assessment, with LCP, INP and CLS each marked good, needs improvement or poor.
  4. Note whether the data describes this page or your whole site. PageSpeed Insights falls back to site wide data when a single page has too few visits.
  5. Scroll down to the performance score and the diagnostics. This is the lab test.
  6. Repeat for your top service page, one location page and the page your Book button opens.

Write the results down with the date. You will want a clear before and after.

See your Surge Score in about a minute

Does ChatGPT recommend your practice?

The Surge Score measures five things about your practice website: how often ChatGPT, Gemini and Perplexity recommend you, your PageSpeed, your structured data, your indexing and your reviews, each with its evidence and date. Free.

Which number actually matters

The pass or fail Core Web Vitals assessment at the top is the number to manage. It passes only when the 75th percentile of real visits is good for all three metrics, which is the closest view you can get of how Google sees the page.

The 0 to 100 performance score is a weighted blend of lab measurements. 90 or above is labeled good, 50 to 89 needs improvement and below 50 is poor. It is handy for spotting problems, and the diagnostics underneath tell your developer where to look. But a page can pass with real users while scoring modestly in the lab, or score well in the lab and still frustrate patients.

If a vendor reports only the lab score, ask for the real user results too. See what Core Web Vitals measure for what each metric means.

When there is not enough real user data

Many practice sites see a message that there is not enough real world data for a page. Google only includes pages and sites with enough visits in the Chrome User Experience Report, and many specialty practices fall below that line for individual pages. PageSpeed Insights then tries to show data for the whole site instead. If the site as a whole also falls short, no real user data appears at all.

That is not a failing grade. In that case, rely on three things: the lab results on the mobile setting, the phone test described below, and your own booking numbers. A page that does well in the slowed down mobile lab test is usually in good shape for real patients.

Search Console: the whole site at once

PageSpeed Insights tests one page at a time. Google Search Console's Core Web Vitals report covers the whole site with the same real user data, grouping similar pages together and reporting mobile and desktop separately.

Each group takes the status of its worst metric, so a group with good LCP and poor CLS shows as poor. Click into a group to see example pages and which metric is failing. After a fix, you can start validation, which monitors the issue over a 28 day period.

If the report says no data is available, your property is new or the site does not have enough traffic. That is common for smaller practices and is not a penalty. Setting up Search Console is covered in Search Console for doctors.

The phone in your pocket test

Tools simulate. Your phone does not. Once a month, try this:

  • Turn off wifi so you are on cellular data
  • Close other apps and open a private browser tab, so nothing is cached
  • Search your practice name on Google and tap your listing
  • Count how long until you see the main photo and heading
  • Tap Book or Call and note whether it responds right away
  • Watch for anything that jumps as the page settles

Then do it again as a new patient searching for a service you offer, starting from Google rather than your homepage. It takes five minutes and catches problems numbers miss, like a pop up that covers the booking button. The mobile versus desktop guide explains why your phone test may feel faster than the lab says.

Why your score changes every time

Run the same lab test three times and you may get three different scores. Lighthouse's documentation says most of that variability comes from changing conditions, such as network routing, ads or tests running on the page, and the device or software running the test, rather than from the tool itself.

So never judge a fix on one run. Run each page three times and keep the middle result, test at similar times of day, and compare like with like: mobile with mobile, the same page before and after. Real user data smooths this out over 28 days, which is one more reason it is the number to manage. For a complete routine, use the one hour speed checklist.

Checklist

  • Test the homepage, top service page, one location page and the booking page on mobile
  • Record pass or fail for the Core Web Vitals assessment, plus LCP, INP and CLS
  • Note whether the data describes the page or the whole site
  • Run each lab test three times and keep the middle result
  • Check Search Console's Core Web Vitals report for mobile groups marked poor
  • Do the phone test on cellular data once a month

Frequently asked

Is another speed tool better than PageSpeed Insights?

Other tools can be useful for diagnosis, but most run lab tests with their own settings and locations. For the question that matters, how real visits to your site perform, PageSpeed Insights and Search Console use Google's own real user data. Start there and use other tools to dig into causes.

Should I test on mobile or desktop?

Both, but mobile first. Google measures them separately, phones are usually slower, and many patients look you up on a phone. If mobile passes, check desktop to confirm.

Why does Search Console say one thing and PageSpeed Insights another?

Check that you are comparing the same thing. Search Console groups similar pages, PageSpeed Insights may show one page or the whole site, and the lab score lower on the PageSpeed Insights page is a different measurement altogether. Compare real user data with real user data, on the same device type.

How often should I test?

Monthly for your key pages, and every time someone adds a widget, changes a template, adds a tracking tag or launches a redesign. Speed problems usually arrive with changes, so testing after each change catches them early.

Sources

  1. About PageSpeed Insights, Google for Developers
  2. Why lab and field data can be different (and what to do about it), web.dev (Google)
  3. Core Web Vitals report, Search Console Help (Google)
  4. CrUX methodology, Chrome for Developers
  5. Lighthouse performance scoring, Chrome for Developers
Prefer to talk it through?

Book a strategy call with the team.

Twenty minutes. We'll walk through the specific opportunities in your market and what a MedPracticeGrowth engagement would look like for your practice.

AI patient discovery audit

Is AI recommending you, or your competitor?

See who ChatGPT, Gemini and Perplexity recommend to patients in your area.

Free. Built by physicians.