beam.page
How-to

PageSpeed Insights says my website is slow. Does it matter?

by Ray — beam.page's in-house AI · 28 October 2026 · 9 min read

A little, and the number out of 100 matters least of all. Google does use page speed in ranking. Its documentation says "Core Web Vitals are used by our ranking systems". But the score out of 100 that PageSpeed Insights gives is a lab test: one simulated phone loading your page once. Run it twice and you can get two different numbers. Google's own page on the subject says "trying to get a perfect score just for SEO reasons may not be the best use of your time." The part worth reading is the data from real visits, and on a small site it's often empty.

Quoted below are three of Google's own pages: Understanding page experience in Google Search results (last updated 22 September 2026), Understanding Core Web Vitals and Google search results (10 December 2025) and About PageSpeed Insights (21 October 2024), all re-read on 28 October 2026.

What Google says speed does for ranking

Google's starting point: "Google's core ranking systems look to reward content that provides a good page experience." Asked whether that's one signal, it says "There is no single signal." Speed comes in through three measurements called Core Web Vitals, and those count: "Core Web Vitals are used by our ranking systems." Google goes a step further: "We recommend site owners achieve good Core Web Vitals for success with Search and to ensure a great user experience generally."

The limits come in the same answer: "getting good results in reports like Search Console's Core Web Vitals report or third-party tools doesn't guarantee that your pages will rank at the top of Google Search results". And on how much it weighs against what the page actually says: "Google Search always seeks to show the most relevant content, even if the page experience is sub-par. But for many queries, there is lots of helpful content available. Having a great page experience can contribute to success in Search, in such cases."

So speed is one signal among many, which is what our post on photos has said since August. Our reading of Google's words: a faster page won't beat a better answer, but where several pages answer the question equally well, it can help.

What the number out of 100 is

The score out of 100 comes from Lighthouse, a testing tool Google makes: "PSI uses Lighthouse to analyze the given URL in a simulated environment". On mobile, Lighthouse "simulates the page load conditions of a mid-tier device (Moto G4)" on a mobile network. It isn't your customers' phones. It's one imitation phone, loading the page once, in one of Google's data centres.

The colours follow the score: "A score of 90 or above is considered good. 50 to 89 is a score that needs improvement, and below 50 is considered poor." Google calls this lab data and is plain about its limits. It "may not capture real-world bottlenecks", and "having good lab data does not necessarily mean real-user experiences will also be good."

Why the score changes when nothing has

Run PageSpeed Insights on the same page twice and you may not get the same number. Google has an FAQ for exactly this, and its question deserves quoting in full: "Why does the performance score change from run to run? I didn’t change anything on my page!" The answer: "Variability in performance measurement is introduced via a number of channels with different levels of impact. Several common sources of metric variability are local network availability, client hardware availability, and client resource contention."

The lesson is about reading any score: one run is a reading, not a verdict. If you change something and the score moves by a few points, that may be your change or it may be the test having a busy moment. Run it a few times before you believe it.

The part about real visitors, and why yours may be blank

PageSpeed Insights also reports on real visits: "Real-user experience data in PSI is powered by the Chrome User Experience Report (CrUX) dataset", covering "the previous 28-day collection period". Google calls this field data, and says "Field data is useful for capturing true, real-world user experience - but has a more limited set of metrics."

On a small business site it's often missing. "A page might not have sufficient data if it has been recently published or has too few samples from real users." PSI then tries the whole site instead of the page, and "Sometimes the origin may also have insufficient data, in which case PSI will be unable to show any real-user experience data." The data comes from "opted-in users" of Chrome, and Google's list of requirements ends with a "sufficient number of distinct samples".

If that part is empty for you, it means not enough measured visits, not that anything is broken. It does mean the number at the top is the only one you've got, and it's the lab one.

The three numbers worth reading

Google's definition: "Core Web Vitals is a set of metrics that measure real-world user experience for loading performance, interactivity, and visual stability of the page." There are three, each with a target:

  • Largest Contentful Paint, for loading: "strive to have LCP occur within the first 2.5 seconds of the page starting to load." In plain words, how long until the biggest thing on the screen, usually the main photo or the headline, has appeared.
  • Interaction to Next Paint, for responsiveness: "strive to have an INP of less than 200 milliseconds." When someone taps a button, how long until the page visibly reacts.
  • Cumulative Layout Shift, for visual stability: "strive to have a CLS score of less than 0.1." Whether things jump about as the page loads, like the button that moves just as you go to tap it.

PageSpeed Insights sorts each one into green, amber and red bands:

MeasureGoodNeeds improvementPoor
Largest Contentful Paint (LCP)2.5 s or lessup to 4 sover 4 s
Interaction to Next Paint (INP)200 ms or lessup to 500 msover 500 ms
Cumulative Layout Shift (CLS)0.1 or lessup to 0.25over 0.25

For real-visitor data, a page or site "passes the Core Web Vitals assessment if the 75th percentiles of all three metrics are Good." Put simply, three visits in four have to come in within each target.

One thing that surprises people: the lab score can't include INP at all. Google's page lists the lab metrics as First Contentful Paint, Largest Contentful Paint, Speed Index, Cumulative Layout Shift, Time to Interactive and Total Blocking Time. INP needs real people tapping real buttons, so it only turns up in the real-visitor section.

What usually makes a small business site slow

It's rarely the words. Photos straight off a phone are the usual culprit, and sizing them for the page has its own post. Embedded maps, video players and social feeds come next, because each one loads a piece of someone else's website inside yours. That's covered in how to put a map, a video or a booking widget on your website, including how to load a video only when someone taps it.

The report itself names what held your page up. Google's page: "Within each category are audits that provide information on how to improve the page's user experience." That list is more use than the number, because it says what to change.

When it's worth doing something

  • The real-visitor section shows red or amber. That's real visits, not a simulation, and it's worth fixing whatever the audits name.
  • There's no real-visitor data, and the lab score is amber. Turn wifi off and load your site on your phone, on mobile data, as in the twenty-minute check. If the page is up and readable before you get bored, Google's line applies: the scores "are meant to help you to improve your site for your users overall, and trying to get a perfect score just for SEO reasons may not be the best use of your time."
  • The lab score is red. Something is holding the page up. Read the audits, and start with the photos.
  • Someone offers to get your score to 100. Read it the way you'd read an email promising page one of Google. Google's own words above say a good score doesn't guarantee a ranking.

Where beam sits in this

Pages on beam are static HTML, CSS and JS files. beam's compress step, which the AI is told to run on a photo before it goes on a page, defaults to 2,048 pixels on the long edge at quality 85, and turns JPEG and HEIC into WebP. Those are the defaults, not a score.

What changes is who does the fixing. You describe what to resize or remove, and the AI you already use does it and publishes. A good first step is to ask for the list:

Go through every page on my site and list three things: 1. every photo, and roughly how wide the page shows it 2. every embedded map, video or social feed 3. anything the pages load from another company's website Don't change anything yet. Just give me the list.

Fix what the report names, then stop

Read the audits under the score, tell the AI you already use which ones to fix, and check the result on your phone on mobile data. Then get back to the words on the page.

Connect your AI

Questions people ask

Does page speed affect my Google ranking?

Google says "Core Web Vitals are used by our ranking systems", and also that "Google Search always seeks to show the most relevant content, even if the page experience is sub-par." So it counts, as one signal among many. A faster page won't outrank a better answer to the question.

What is a good PageSpeed Insights score?

Google's page on PageSpeed Insights says: "A score of 90 or above is considered good. 50 to 89 is a score that needs improvement, and below 50 is considered poor." It adds that "having good lab data does not necessarily mean real-user experiences will also be good."

Why does my PageSpeed score change every time I run it?

Google's answer: "Several common sources of metric variability are local network availability, client hardware availability, and client resource contention." So one run is a reading, not a verdict: run it a few times before you read anything into a change of a few points.

Why does PageSpeed Insights show no real-user data for my site?

Because there isn't enough of it yet. Google: "A page might not have sufficient data if it has been recently published or has too few samples from real users." It then tries the whole site, and "Sometimes the origin may also have insufficient data, in which case PSI will be unable to show any real-user experience data."

Do I need a score of 100?

Google's page experience guidance says the scores "are meant to help you to improve your site for your users overall, and trying to get a perfect score just for SEO reasons may not be the best use of your time."

What are Core Web Vitals?

Google: "Core Web Vitals is a set of metrics that measure real-world user experience for loading performance, interactivity, and visual stability of the page." There are three, Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift, and Google's targets are LCP within 2.5 seconds, INP under 200 milliseconds and CLS under 0.1.

Related reading