Speed

Why your website is slow

What actually makes a page slow on a phone. The tests, the numbers, and what to fix first.

Most slow sites are slow after they were built heavy. The page hands a phone more to download and more to run than the job needs, so the person waiting stares at a blank screen. It comes down to three things: how many bytes the page sends, how much work the phone does after those bytes land, and how far away the server sits. The first two explain most of it.

Most pages are carrying too much

The HTTP Archive Web Almanac measured the web for its 2025 Page Weight chapter, published January 15, 2026. The median mobile home page used 22 KB of HTML, 77 KB of CSS, 122 KB of fonts, 632 KB of JavaScript, and 911 KB of images. The median page came to 2.56 MB on mobile and 2.86 MB on desktop.

Images are the biggest single part, and the easiest part to fix. A jobsite photo often gets uploaded at full camera size, then squeezed into a small box on screen. The visitor still downloads the whole file.

Slow to show up and slow to respond are two different problems

Google's web.dev page on Largest Contentful Paint, last updated September 4, 2025, calls 2.5 seconds or less good and anything over 4 seconds poor. That is how long the main thing on screen takes to appear.

A site can pass that and still feel broken. Google's page on Interaction to Next Paint, last updated September 2, 2025, calls 200 milliseconds or less good and anything over 500 milliseconds poor. That is the wait after someone taps. Chat bubbles, review widgets, and tracking scripts keep a phone busy long after the page looks finished.

A speed score is not what your customers felt

Those numbers are graded at the 75th percentile of visits, so one visitor in four can be slower and the page still passes. Real visitor data comes from the Chrome User Experience Report, and its methodology page from Chrome for Developers, last updated June 20, 2024, says a page has to be publicly discoverable and popular enough to qualify. The exact number is not disclosed. A small local site often never clears that bar, so the test falls back to a lab run on a simulated phone.

This site scores 100 on Google's PageSpeed test on desktop and mobile. That is still a lab score, and it only shows the build is light. It says nothing about a customer standing in a basement with one bar of signal, and a faster site is not a promise of more calls. If the site already shows up in about two seconds on a phone on cell data, speed is not the problem, and nobody should sell a rebuild over it.

What to do next

Work in this order and re-test after each step, so you can tell which change helped.

  • Run PageSpeed Insights on the page customers actually land on, not just the home page.
  • Check whether real visitor data appears. If it does not, treat the score as a lab guess.
  • Find the heaviest images on that page and save them at the size they are shown at.
  • Count the outside scripts on the page. Drop the ones nobody uses.
  • Test three times on the same connection and keep the middle score.

Everything on speed

13 notes
01

6 MIN

Does a live chat widget slow down your website?

Measured chat widget sizes, what they cost in main thread time, and how to check the one sitting on your own site.

Read Does a live chat widget slow down your website?
02

7 MIN

Does server location affect local SEO? What Google says

Where your website is hosted does not decide which town ranks. Here is what Google publishes, the latency math, and when a host move is worth it.

Read Does server location affect local SEO? What Google says
03

6 MIN

Why your jobsite photos slow your website down

A photo off your phone is thousands of pixels wide. Your page shows it a few hundred. How to find the heavy ones, what to resize, and what it will not fix.

Read Why your jobsite photos slow your website down
04

7 MIN

Why PageSpeed Insights shows no field data for your site

PageSpeed Insights shows no real user data until enough Chrome visitors reach your page. What the 28 day window needs, and why an empty box is not a fault.

Read Why PageSpeed Insights shows no field data for your site
05

6 MIN

Why GTmetrix and PageSpeed Insights disagree

GTmetrix defaults to a Seattle desktop with no speed limit. PageSpeed Insights fakes a mid-range phone. Same page, two numbers, one to watch.

Read Why GTmetrix and PageSpeed Insights disagree
06

6 MIN

Do you lose half your visitors after three seconds?

The 53% figure is real, and older and narrower than it sounds. What Google and Deloitte actually measured, and the speed target to use now.

Read Do you lose half your visitors after three seconds?
07

7 MIN

Do custom fonts make your website slower

Yes, a little. Here is what a web font really costs on a phone, why it can hide your text, and the setting that decides what shows.

Read Do custom fonts make your website slower
08

6 MIN

Why your page jumps just as someone taps your number

Your page moves while it loads and the tap lands wrong. Here is what causes the jump, the score that measures it, and how to stop it.

Read Why your page jumps just as someone taps your number
09

7 MIN

What actually makes a website load faster, in order

The four parts of a page load, ranked. What Google's own docs say is usually slow, what rarely is, and how to check your own site.

Read What actually makes a website load faster, in order
10

6 MIN

Interaction to Next Paint: why a loaded page still feels slow

What Interaction to Next Paint measures, the 200 millisecond mark Google calls good, and how to check your own number on a phone.

Read Interaction to Next Paint: why a loaded page still feels slow
11

5 MIN

HVAC website slow in summer: what breaks under a heatwave

Heatwave week is when HVAC sites buckle. What runs out on a shared host, why static files do not, and how to test your own site in two minutes.

Read HVAC website slow in summer: what breaks under a heatwave
12

5 MIN

Plumber website design that loads fast on a phone

A slow plumbing site loses the call. Here is the speed target, what slows a page down, and a check you can run on your own phone.

Read Plumber website design that loads fast on a phone
13

2 MIN

This site was built in a day. Here is what it scored.

It launched with nine pages and scored 100 on Google PageSpeed for phone and desktop. Here is how, and what has changed since.

Read This site was built in a day. Here is what it scored.

Want this run on your business?

A free, one-time search check. A real person looks at what Google and ChatGPT say about you. You get an email back with what to fix first.

Your next step

Let's plan your website.

Tell us about your business. We'll get back to you to set up a call.

Add business details (optional)
How we use your details

Prefer email? go@galloping.ai