How to read your Google PageSpeed score without panicking

A PageSpeed score is Google's lab test of one page, run on a simulated slow phone, boiled down to a number out of 100. It is not a grade on your business, it is not what your customers experience, and a 41 is a thermometer reading, not a diagnosis. Here is how to read the report so you fix the right thing and ignore the rest.

What the number is

The performance score is "a weighted average of the metric scores" (Source: Chrome for Developers). Five measurements go in, each with a weight: Largest Contentful Paint 25 percent, Total Blocking Time 30 percent, Cumulative Layout Shift 25 percent, First Contentful Paint 10 percent, Speed Index 10 percent (Source: Chrome for Developers, Lighthouse 10 weights). Google's color bands are 0 to 49 red, 50 to 89 orange, 90 to 100 green (Source: Chrome for Developers).

Nobody has ever hired a plumber because of his Total Blocking Time. The score is a composite for engineers. You want the one metric underneath it that describes what a person sees.

What "lab" means, because it changes everything

The test does not run on your phone or your customer's phone. PageSpeed Insights states that "Lighthouse simulates the page load conditions of a mid-tier device (Moto G4) device on a mobile network for mobile" (Source: PageSpeed Insights). The default simulation is a connection Lighthouse labels Slow 4G, with 150 milliseconds of latency, 1.6 megabits per second down, and a constant 4x CPU slowdown (Source: Lighthouse throttling documentation).

So when your site "scores 38," what happened is: a mid-tier phone of Google's choosing, on a deliberately bad connection, loaded your page once under conditions chosen to expose problems. That is useful. It is also why your own phone on your own Wi-Fi tells you nothing about the score.

The report may also show a field section. That one comes from real Chrome users: "Real-user experience data in PSI is powered by the Chrome User Experience Report (CrUX) dataset." (Source: PageSpeed Insights.) Google explains the difference: "The field data is a historical report about how a particular URL has performed... The lab data is based on a simulated load of a page on a single device and fixed set of network conditions." (Source: PageSpeed Insights.) If there is no field section, everything on your report is lab.

The one number to read first

Largest Contentful Paint, LCP. It "reports the render time of the largest image, text block, or video visible in the viewport, relative to when the user first navigated to the page" (Source: web.dev). On a trade site that is almost always the hero photo. Google's line: "sites should strive to have Largest Contentful Paint of 2.5 seconds or less" (Source: web.dev).

Find LCP on the report. If it is at or under 2.5 seconds in the lab, your main image appears quickly on a slow phone, and the rest of the score is detail. If it is well over, that is the thing to fix, and it is usually the hero image being far larger than the screen it lands on.

What we measured locally

In September 2026 we ran Google's lab test across Capital Region home-service sites. Lighthouse mobile main-image time, LCP: 102 sites measured, median 7.75 seconds, 98 over Google's 2.5-second line, 37 over 10 seconds. Page weight: 102 sites, median 3.3 MB, 55 over 3 MB. (Source: DarkTiger Studios gate, September 2026.) The lab test runs on a simulated throttled phone, not real visitor data, and the sample is Capital Region home-service businesses from chamber directories and license rolls, not a random sample.

Page weight is a measurement, not the cause of the wait. A heavy page can show its main image fast if the image is sized and the rest loads later. A light page can still be slow if the image is blocked behind scripts. Read weight as a clue, read LCP as the result.

What not to panic about

The score moves between runs. Google lists reasons: changes in ads being served, internet traffic routing changes, testing on different devices, browser extensions, antivirus software, and says "it might be more useful to think of your site performance as a distribution of scores, rather than a single number" (Source: Chrome for Developers). Run it three times. If you get 38, 44 and 41, you have a 41-ish site, not a crisis and not three crises.

And we do not have a number for what a slow lab score costs you in calls. Nobody has measured that for a roofer in Troy. Anyone who quotes you a percentage should be asked where it came from.

Next step

Run PageSpeed Insights on your home page, mobile tab, and write down the LCP. If it is over 2.5 seconds in the lab, the Website Fix compresses and lazy-loads the images, repairs dead links, makes the phone number tappable, forces one secure address, and re-runs the gate with a before-and-after report, for $700. Details at /services#fix.