← All owner guides

Owner guide · 11 min read

How Fast Should My Website Load?

Your main content should show up within 2.5 seconds on a phone. That is Google's mark for good. Most Utah trade sites miss it: in our October 2026 scan of 404 of them, the typical site took 8.4 seconds and only 4% made 2.5. Heavy pages and add-on scripts are the usual causes.

KC Hansen

Published October 27, 2026

2.5 s

Google's mark for a good main-content load time on a phone

8.4 s

the typical Utah trade site in our scan of 404 sites (Oct 2026)

4%

of those 404 sites meet Google's 2.5-second mark

Your website’s main content should appear within 2.5 seconds on a phone. That’s the line Google draws between good and needs work. Most local trade websites are nowhere near it. In October 2026 we tested 404 Utah trade business websites on Google’s own phone test, and the typical one took 8.4 seconds to show its main content. Only 17 of them, about 4%, made it in 2.5 seconds or less.

What does “loaded” actually mean?

“Loaded” means the visitor can see and use the page, not that every last file has arrived. Google measures it with three numbers, called Core Web Vitals, and publishes a good target for each (web.dev) and a poor line for each (web.dev):

What Google measuresIn plain wordsGoodPoor
Largest Contentful PaintHow long until the main picture or headline shows2.5 seconds or lessover 4 seconds
Interaction to Next PaintHow long the page takes to react when you tap200 milliseconds or lessover 500 milliseconds
Cumulative Layout ShiftHow much the page jumps around while it loads0.1 or lessover 0.25

The first one is the one owners feel. It’s the gap between tapping your link on Google and seeing anything worth looking at. If someone is standing in a flooded basement, that gap is where they hit the back button and call the next plumber.

Google judges these at the 75th percentile of real visits, which means three out of four visits should meet the target, not just your best one (web.dev).

How fast are Utah trade websites?

Most are slow on a phone, and many are very slow. We scanned the websites of 409 independent Utah trade businesses, the ones Google shows in its map results when you search for roofers, plumbers, HVAC, electricians, landscapers, painters, cleaners, garage door, pest control and concrete companies in Provo, Salt Lake City, Layton, Lehi and Sandy. Each site was run through Google Lighthouse, the engine behind PageSpeed Insights, three times on phone settings. 404 of them gave a speed result.

How long Utah trade websites take to show their main content on a phone

Share of 404 sites in each time band. Google's "good" line is 2.5 seconds.

2.5 s or less (good)4% · 17 sites
2.5 to 4 s13% · 51
4 to 6 s18% · 73
6 to 10 s23% · 94
10 to 20 s28% · 114
Over 20 s14% · 55
Source: YourProject.ai scan of 404 independent Utah trade business websites, October 21-22, 2026. Google Lighthouse, phone settings, median of three runs. Method at the end of this guide.

The headline numbers from the same scan:

  • 8.4 seconds is the median time to show the main content. Half the sites were faster, half slower.
  • 58 out of 100 is the median Google speed score on a phone. 30% of sites scored under 50, and only 6% scored 90 or better.
  • 83% took longer than 4 seconds, which Google calls poor.
  • 42% made you wait for a full page of text or photos for 10 seconds or more.

These are lab numbers: a simulated mid-range phone on a slow mobile connection, the same for every site. A real visitor on strong 5G will often see a faster page. But Google’s test is the same ruler for everyone, and the gaps between sites are real.

Tip: the speed score swings a few points from one run to the next. Run PageSpeed Insights three times and use the middle result, which is what we did for every site in our scan.

What slows a website down?

Mostly the size of the page and the add-ons loaded with it. The server is rarely the problem. In our scan, 94% of sites had a server that answered in under 0.8 seconds (the median was 0.07 seconds). The waiting happened after that, while the phone downloaded and ran everything the page asked for.

Page weight. The median Utah trade homepage weighed 2.9 MB, and only 23% came in under 1.5 MB. Weight lines up closely with waiting time:

Heavier pages take longer to show up

Median main-content time on a phone, by total homepage size. 404 Utah trade sites.

Under 1.5 MB (92 sites)4.3 s
1.5 to 3 MB (120 sites)8.1 s
3 to 6 MB (103 sites)10.4 s
Over 6 MB (89 sites)14.9 s
Source: YourProject.ai scan, October 2026. Medians; sites with a speed result only. A pattern, not proof that weight alone causes the delay.

Photos are the usual culprit. They made up about 0.8 MB of the typical page, and only about a third of images used modern compressed formats (WebP or AVIF). A single photo straight from a phone camera can be 3 to 5 MB on its own.

Add-on scripts. Chat bubbles, review widgets, booking pop-ups, video players, heat-map tools and ad trackers each load code from another company. The typical site loaded code from 6 outside companies. Sites with 10 to 19 took a median 10.3 seconds to show their main content and froze for about 0.7 seconds while that code ran. Sites with 4 or fewer took 5.8 seconds. Only 40% of all sites passed Google’s “not frozen” check, which asks that the page stay responsive for all but 200 milliseconds while loading.

What the site is built on matters less than what’s piled on it. Here are the three builders with enough sites to compare fairly:

BuilderSitesMedian speed scoreMedian main-content timeMedian page weight
WordPress148569.6 s3.0 MB
Custom or other81657.9 s2.7 MB
GoHighLevel33539.8 s5.0 MB

Wix (29 sites), GoDaddy (24) and Squarespace (17) had too few sites in our sample to rank fairly, so we don’t. Every builder had fast sites and slow sites.

Do ads and big review counts make sites slower?

They go together with slower sites in our data. This is a pattern, not proof of cause. Sites with Google Ads or Meta (Facebook) ad tracking installed had a median speed score of 49 and took 10.6 seconds to show their main content. Sites without them scored 64 and took 7.3 seconds. The ad-tracking sites were also nearly twice as heavy (3.9 MB against 2.1 MB) and loaded code from more outside companies (10 against 4).

The businesses with the most Google reviews had the slowest sites of all. The 32 businesses with 1,000 or more reviews had a median speed score of 45 and a main-content time of 15.8 seconds. Bigger companies tend to add more: chat, call tracking, booking, retargeting. Each one costs time on a phone.

If you pay for clicks, this is the expensive version of the problem. You pay for the visit, then the page makes the visitor wait.

Does speed affect your Google ranking?

A little, and less than most people think. Google says plainly that “Core Web Vitals are used by our ranking systems,” and in the same guide says good scores don’t guarantee a top spot: the most helpful, relevant page still comes first (Google Search Central). Speed won’t rescue thin pages, and it won’t sink a page that answers the question better than anyone else.

Where speed matters most is after the click. A visitor who waits 10 seconds on a blank screen may never see your phone number. The best-known study on this was commissioned by Google and run by Deloitte in 2020 on the mobile sites of large retail, travel, luxury and lead-generation brands: when every speed measure on a mobile site improved by just a tenth of a second, retail conversions rose 8.4% and lead-generation sites saw 21.6% more people move from one step of the form to the next (web.dev, Milliseconds Make Millions). Those are large national brands, not local trades, so treat it as a direction, not a forecast. Our own data can’t show how many calls a slow site costs any one business. Nobody can know that without the business’s own call records. But making customers wait is never the reason someone picks you.

How do you test your own site?

You can do it in ten minutes with free tools.

  1. Open PageSpeed Insights and paste your homepage address. Stay on the Mobile tab.
  2. Look for “Largest Contentful Paint” in the diagnostics. Under 2.5 seconds is good. Over 4 seconds is poor.
  3. Run it three times and use the middle result.
  4. Check the field data box at the top. If your site has enough visitors, Google shows how fast it really was for real Chrome users over the past 28 days. Small sites often don’t have enough visits to show this, and that’s normal.
  5. Scroll to “Avoid enormous network payloads” to see your page weight and the biggest files.
  6. Test on your own phone on cell data. Turn off Wi-Fi, search your business on Google, and tap your link. Count the seconds until you could call.

What fixes make the biggest difference?

Start with photos and add-ons. They are the cheapest wins and the biggest ones in our data.

FixWhy it helpsEffort
Resize and compress photos (WebP, under about 200 KB each)Photos are the biggest part of most pagesLow
Load lower photos only when someone scrolls to themThe first screen shows soonerLow
Remove add-ons nobody uses (old chat, unused trackers, sliders)Fewer outside companies, less freezingLow
Replace an auto-playing video header with one strong photoVideo is usually the heaviest thing on the pageLow to medium
Load the remaining trackers after the page showsKeeps them from blocking the first screenMedium
Rebuild a heavy theme or page builderSome themes ship megabytes of code you don’t useHigh

What happened when we rebuilt 18 sites?

They got lighter and faster, and we aren’t at Google’s line on every one yet. We built new sites for 18 businesses that already had a live website and tested the old and new versions with the same test on the same day:

  • Speed score: median 55 before, 83.5 after. Better on 15 of the 18.
  • Main content on a phone: median 9.5 seconds before, 4.5 seconds after. Faster on 12 of the 18.
  • Page weight: median 3.1 MB before, 0.7 MB after. Lighter on 14 of the 18.

Three of the old sites were already quick, and our new versions didn’t beat them. Across all 23 sites we’ve built, the median main-content time is 4.0 seconds and 4 of 23 make Google’s 2.5-second mark.* That is a lot better than the Utah median of 8.4, and still not where we want it. Our large opening photos are the main reason, and they are what we are working on.

*Our numbers here are from preview copies of each site, tested before launch. A preview doesn’t have the setup a launched site gets: a delivery network that keeps copies of every page close to the visitor, and long-term caching so repeat files aren’t downloaded again. The old sites were tested as they run today, with whatever setup they already had. We expect our launched sites to be faster than these preview numbers, and we’ll publish the launched-site numbers when we rescan the same sites.

How we handle speed for you

Every site we build is small and light: photos are resized for phones and load as you scroll, and there are no ad trackers. Our best build so far scored 100 in all four categories on Google’s speed test. Sites are hosted on Amazon Web Services with HTTPS and security headers.

Speed checks are part of Care, our $99-a-month plan, along with broken-link checks. On Growth, every month we also read Google’s report of the searches that found you and fix the pages that fall short. The build is $5,000, one time, and you see your new site working on your own phone before you pay anything. We build everything Google needs. Nobody can promise a ranking.

See what we build and what it costs, look at sites we’ve built, or apply in about two minutes. If your site has other phone problems, start with does my website need to work on phones?, and if you’re wondering whether your host is to blame, read does it matter where my site is hosted?.

How we measured

  • Who: the websites of 409 independent Utah trade businesses, taken in order from Google’s local map results for 10 trades in Provo, Salt Lake City and Layton (plus Lehi and Sandy for some trades). National franchise sites were left out of these numbers. Sites that were broken, parked or blocked automated visits weren’t scored.
  • When: October 21-22, 2026.
  • How: Google Lighthouse on phone settings, three runs per site, middle result used, all sites in one shuffled queue on one machine so no group got a better time of day. 5 sites gave no speed result and are left out of the speed numbers.
  • Limits: these are lab tests, not real-visitor data. Google’s map results put stronger, better-reviewed businesses first, so if anything this sample is better than the average Utah trade site. Patterns here are correlations, not causes.
  • Who we are: we build websites for local businesses, so we have a stake in this topic. That’s why the method is here for anyone to check. We don’t name any business.

Sites we've built

Real sites we built for local businesses. See all our work →

Speed check for your site

0 / 6

Tap to tick. Your ticks stay on this device.

Questions

Asked constantly. Answered honestly.

Is the PageSpeed Insights score the number that matters? +

It's a summary, not the goal. Look at the main-content time (Largest Contentful Paint) on the mobile tab first. A site that shows its main content in under 2.5 seconds is in good shape even if the score is in the 80s. The score also moves a few points between runs, so test three times.

My site loads fast on my computer. Why does the test say it's slow? +

Your computer is on fast office internet with a fast processor, and your browser has already saved most of the site. The test pretends to be a mid-range phone on a slow mobile connection, visiting for the first time. That's closer to a new customer standing in a driveway.

Will a faster site rank higher on Google? +

Speed is one of many signals, and Google says the most relevant page still wins. A fast page won't beat a better answer. But a slow page can lose people before they read anything, and that costs you calls whatever your position.

Do I need to switch hosting to get faster? +

Usually not. In our scan, 94% of Utah trade sites had a server that answered quickly. The slow part was what the server sent: big photos and outside scripts. Fix those first.

$5,000 for your website + from $99/mo after. See it before you pay.

See your new site free
KC Hansen

founded YourProject.ai in Lehi, where we build websites for local businesses. His whole career in IT (security, cloud, DevOps) means every site is fast, secure and built right, and kept up to date every month. More about KC and every guide →