Home/Guides/Page speed
Care guide

Page speed and Core Web Vitals,
in plain English.

Core Web Vitals are three measurements Google uses to judge how a page feels to real visitors: how fast the main content shows up, how quickly the page responds when someone taps, and whether things jump around while it loads. Here's what each one means, what counts as good, why your phone score is the one that matters, and which fixes are worth doing first.

Updated November 2026 · 12 min read

A phone in a dashboard mount, showing a map
Photo: Malte Luk / Pexels

The short answer

Page speed is how quickly your site loads and responds for the person holding the phone. Google measures it with three Core Web Vitals:

  • LCP: does the main content show up fast? Good is within 2.5 seconds.
  • INP: does the page respond quickly when tapped? Good is 200 milliseconds or less.
  • CLS: does the page hold still while it loads? Good is a score of 0.1 or less.

Each is judged at the 75th percentile of real visits, on mobile and desktop separately, and a page passes when it meets all three. On most small business sites, the problems come from the same short list: oversized photos, too many scripts, chat widgets and sliders. The fixes are usually cheaper than people expect, and images come first.

Speed matters to Google, but as one signal among many. It matters more to your customers, who leave slow pages before they ever see your phone number.

The three Core Web Vitals, one at a time

Here's what each metric actually measures, what Google counts as good, and what usually causes trouble.

MetricWhat it measuresGoodCommon causes when it's bad
LCP (Largest Contentful Paint)How long until the biggest thing on screen (usually the main photo or headline) has loadedWithin 2.5 secondsA huge hero photo, slow hosting, a slider, web fonts holding up the text, scripts that load first
INP (Interaction to Next Paint)How quickly the page visibly responds when someone taps, clicks or types200 milliseconds or lessToo many scripts: tracking tags, chat widgets, pop-ups, heavy page builders
CLS (Cumulative Layout Shift)How much the layout jumps around while the page loads0.1 or lessImages without set sizes, ads and embeds that load late, banners that push content down, font swaps

Thresholds from Google's web.dev Core Web Vitals article, checked November 2026. Each is measured at the 75th percentile of page loads, on mobile and desktop separately. A page passes when it meets all three.

LCP: is the main thing there yet?

Largest Contentful Paint is the moment the biggest element in view finishes loading. On most business sites that's the hero photo at the top of the page, or a large block of headline text. It's the closest single number to "has the page loaded?" from the visitor's point of view. Slow LCP almost always means a heavy image, a slow server, or something loading in front of the main content.

INP: does it respond when I tap?

Interaction to Next Paint measures the delay between a tap, click or key press and the page visibly reacting, across the whole visit. It replaced an older measure, First Input Delay, in 2024. A page can look loaded and still feel stuck, because scripts are busy in the background. The menu button that does nothing for a moment, the form field that lags: that's INP. The cause is nearly always too much JavaScript, often from third-party tools.

CLS: does it hold still?

Cumulative Layout Shift scores how much visible content moves unexpectedly. You've felt it: you go to tap a link and the page jumps because an image or banner loaded above it, and you tap an ad instead. The score has no units; lower is better, and 0.1 or less is good. The usual fix is simple: give images and embeds a fixed size so the page saves room for them.

The 75th percentile rule

Google doesn't judge your page on one test, or on an average. It looks at real page loads and takes the 75th percentile: the experience that three out of four visits meet or beat. Mobile and desktop are judged separately.

Example: how the 75th percentile works

Say your service page had 100 visits from phones last month. Line up their LCP times from fastest to slowest. The 75th one in the line is your score. If that visit saw the main photo load in 2.3 seconds, the page passes LCP on mobile. If it took 3.1 seconds, it doesn't, even if most visitors on fast office wifi saw it in under a second.

That's the point of the rule. It's set so the page has to work for the slower phones, the weaker signal out past town, and the older handsets, not just for you on your desk.

To pass overall, a page needs all three metrics in the good range at the 75th percentile. Two out of three isn't a pass.

Field data vs lab data

When you test a page, you'll see two kinds of numbers, and mixing them up causes a lot of confusion.

  • Field data comes from real visitors using Chrome, collected over recent weeks. This is what the Core Web Vitals assessment is based on. It's the truth, but it needs enough traffic: many small business pages don't have enough visits to show field data at all.
  • Lab data is a single test run by a tool, on a simulated phone and connection. It's what produces the familiar score out of 100. It's useful for finding problems and checking fixes, but it varies from run to run and isn't what Google uses for the assessment.

So a lab score of 62 doesn't mean you're failing, and a 95 doesn't guarantee you're passing. Use lab tests to diagnose and compare before and after a change. Use field data, where you have it, to know where you actually stand. If your site doesn't have field data yet, run lab tests on a phone setting, several times, and look at the trend rather than one number.

Why your phone score is the one that matters

Desktop scores flatter almost every site. Office computers are fast and office internet is quick. Phones have slower processors, they're often on mobile data, and they're what most local customers use to find a contractor, check a menu or call a clinic.

Google also indexes the mobile version of your site first, so the phone experience is the one it looks at. That's why every test in this guide should be run on the mobile setting, and why the 75th percentile on mobile is the number to watch.

Try it yourself: open your site on your phone on mobile data, standing in a parking lot, and count until the main photo and your phone number are on screen. That one test tells you more than most reports.

What usually slows small business sites down

We look at a lot of small business sites. The same causes come up again and again:

  • Huge images. Photos uploaded straight from a camera or phone, many times larger than the screen showing them. This is the most common cause of slow LCP by a distance.
  • Too many scripts. Analytics, ad pixels, heat maps, review widgets, social feeds. Each is small; together they slow every page and hurt INP.
  • Chat widgets. Many load a lot of code on every page for a button most visitors never press.
  • Sliders and carousels. Several large images loaded to show one at a time, plus movement that can count against CLS.
  • Web fonts. Several families in several weights, loaded from another server, holding up the text.
  • Cheap or overloaded hosting. A server that takes a long time to answer delays everything after it.
  • Page builders. Drag-and-drop builders on WordPress and elsewhere often load the code for every feature on every page, whether it's used or not.
  • Embeds. Maps, video players and booking tools loaded at the top of the page instead of when needed.
A camera, memory cards and a laptop showing a photo being edited
Photos straight off a camera are often many times larger than the screen they're shown on. Photo: Leeloo The First / Pexels

Notice what's not on the list: anything exotic. Most slow sites are slow for ordinary reasons, which means ordinary fixes work.

Fixes, in order of payoff

Do these roughly in this order. The ones at the top are cheap and fix the most.

  1. Fix the images. Resize photos to the size they're shown at (a photo straight off a phone or camera is often several times wider than any screen), save them in a modern format such as WebP, and set their width and height so the page saves room for them. Lazy-load images further down the page, but never the main photo at the top: that one should load first.
  2. Remove scripts you don't use. Old tracking tags, an analytics tool nobody reads, a pixel from an ad campaign that ended. Each one runs on every page load.
  3. Delay the heavy extras. Chat widgets, video players, maps and review feeds can load when someone scrolls to them or taps them, instead of on arrival. A static picture of the map that turns into the real map on tap looks the same and costs almost nothing.
  4. Replace the slider. A rotating carousel loads several big images to show one, and the movement often counts against CLS. One strong photo and a clear headline almost always does the job better.
  5. Trim the fonts. Use one or two font families and only the weights you need, host them with the site, and let the text show in a fallback font while the web font loads.
  6. Reserve space for anything that loads late. Banners, embeds and ads should have a fixed slot so they don't push content down when they appear.
  7. Look at the hosting. If the server itself is slow to answer, every page starts late. Caching and a content delivery network (servers in many cities) help; sometimes a better host is the fix.
  8. Consider the build itself. A page builder that loads a large amount of code on every page can only be tuned so far. If you've done everything above and phones are still slow, the next step is a lighter rebuild.

For the first step, our free Image Optimizer resizes and compresses photos in your browser before you upload them. If you take your own job photos (and you should), our guide on real photos vs stock covers getting them right for the web.

A slow site, worked through

Worked example

Say a family restaurant in Sartell has a site that feels slow on phones. The home page has a five-photo slider of dishes, each straight from the photographer at full size, a chat widget, an online ordering embed, three font families, and four tracking tags, two from ad campaigns that ended last year.

Here's the order we'd tackle it in:

  1. Replace the slider with one photo of the signature dish, resized to the width it's shown at and saved as WebP, with its size set. That alone takes most of the weight off the page and steadies the layout.
  2. Remove the two dead tracking tags.
  3. Load the chat widget and the ordering embed only when someone taps "Chat" or "Order online".
  4. Cut the fonts to one family in two weights.
  5. Test again on the mobile setting, three times, and compare against the numbers from before.

None of this changes how the restaurant looks or what it offers. The menu, hours and order button are the same. They just show up sooner, which is the whole point when someone's deciding where to eat from a car.

This is an example we've set up, not a client's real results, but every item in it is common. If you'd like to see what a slow site might be costing in lost enquiries, our Cost of a Slow Website puts a rough number on it from your own traffic. Restaurants have their own needs on top of speed; our restaurant web design page covers those.

How to test your own site

Three tools cover everything a small business needs:

  1. Our Website Speed Test. A quick read on any page, explained in plain words. Good for a first look.
  2. Google's PageSpeed Insights. Enter a page address. If your page has enough traffic, the top section shows field data from real visitors and whether it passes Core Web Vitals. Below that is a lab test with a score and a list of what to fix. Use the mobile tab.
  3. Google Search Console's Core Web Vitals report. Once your site is verified, this groups your pages into good, needs improvement and poor, for mobile and desktop, based on field data. It's the best view across the whole site, though small sites may not have enough traffic for it to show anything.

A few habits make the results useful. Test your key pages, not just the home page: your top service page and your contact page matter as much. Run each lab test a few times and look at the middle result. Write the numbers down, so after a change you can tell whether it helped. And re-test every quarter, because sites get slower as photos and tools pile up. Our website maintenance checklist puts that on the calendar.

Does speed affect your Google ranking?

Google uses Core Web Vitals as one signal among many. Relevance comes first: a fast page that doesn't answer the search won't outrank a slower one that does. Nobody can honestly promise that fixing speed will move you up a set number of places, and you should be wary of anyone who does.

The clearer payoff is with people. A visitor who waits for a slow page on a phone often just goes back and taps the next result. You never see that lost call in any report. If you're not showing up at all, speed is rarely the main reason; our guide on why you're not showing up on Google covers what usually is.

The sites we build are static files served from Cloudflare's edge in 300+ cities, with no database to wait on and no plugins loading code, which is why they tend to start fast. Our speed and security page explains the setup, and our care plans include speed checks and tuning so they stay that way as photos and pages are added. If your current site is slow and the fixes above don't get it there, a redesign may be the cheaper path. We work with businesses from St. Michael to St. Cloud; send us your address and we'll tell you which.

Want a straight answer for your business?

Every Stonekept build is custom-coded, hosted on Cloudflare and kept on a monthly plan. Prices are public: builds from $2,200 with care from $199/mo. See pricing or tell us about your project and we'll give you an honest read, even if that's not us.

Questions

Quick
answers.

What are Core Web Vitals?

Three measurements Google uses for how a page feels to real visitors: Largest Contentful Paint (loading), Interaction to Next Paint (responsiveness) and Cumulative Layout Shift (visual stability).

What is a good Core Web Vitals score?

LCP within 2.5 seconds, INP of 200 milliseconds or less, and CLS of 0.1 or less, each measured at the 75th percentile of page loads. A page passes when it meets all three, and mobile and desktop are judged separately.

Why is my PageSpeed score different every time I test?

The score comes from a lab test, a single simulated run that varies with network and server conditions. Run it a few times and use the middle result. The field data section, when your page has enough traffic, is more stable.

Why is my mobile score so much lower than desktop?

Phones have slower processors and often slower connections, so the same page takes longer to load and respond. Mobile is also what Google looks at first, so it's the score to focus on.

Does page speed affect SEO?

Core Web Vitals are one signal among many in Google's ranking, and relevance matters more. The bigger effect is on visitors: slow pages lose people before they call or fill in a form.

What's the fastest way to speed up my website?

On most small business sites, fix the images first: resize them to the size they're shown, use a modern format like WebP, and set their dimensions. Then remove unused scripts and delay chat widgets and embeds.

Why doesn't my site show field data?

Field data needs a certain amount of real Chrome traffic to a page. Many small business pages don't have enough yet. Use lab tests on the mobile setting until it appears.

Let's build something lasting

A site that loads
before they give up.

Tell us what you need and we'll map out the right approach on a free discovery call.

StudioBuffalo, Minnesota
ServingTwin Cities west metro