Run your main pages through a free mobile speed test, note the page weight and number of requests, and fix the heaviest item first, usually oversized photographs, then whatever delays the first screen.

Fast at the office

Ask an owner how fast their website is and the answer is usually fine. They checked it at the office, on a fast connection, with the site already stored in the browser.

The customer is in a parking lot or on a bus, using mobile data, and loading the site for the first time. Same page, very different wait.

Free tests show what that customer sits through. Most owners have never run one.

Three free tests

Each of these takes a minute and needs nothing but the page address.

The third is the one that shows what the customer actually sees. Its filmstrip captures the screen at intervals while the page loads, so several seconds of blank white become visible rather than a number in a table.

Reading the numbers

The tests use different words, but a handful of numbers matter most.

Page weight is the total size of everything the page downloads, in megabytes. On mobile data every megabyte takes time, and on some plans it costs the visitor money.

Requests are the separate files the page asks for: images, scripts, stylesheets, fonts, tracking codes. Each one is a round trip, and a phone on a mobile network pays a delay for every trip.

Start render is when anything first appears on the screen.

Time to first byte is how long the server takes to begin answering. If that alone is slow, the problem is on the server rather than in the page.

The score is not the point

PageSpeed Insights gives a score out of 100, and owners tend to fixate on it. The score reflects how many of its rules the page follows, not how long a customer waits.

A page can score well and still be heavy, or score badly over items that cost a fraction of a second. Use the list of rules as a checklist, and the weight and loading time as the measure.

Some of the rules it flags are cheap to fix and always worth doing: compression switched off, images larger than they display, and no browser caching for files that never change.

Where the weight usually is

On most small business sites the answer is photographs. A picture straight off a phone or camera is several megabytes and thousands of pixels wide, shown in a space a few hundred pixels wide on a phone.

Slideshows multiply the problem. Five large images rotating across the top of the home page can outweigh the rest of the site, and most visitors only ever see the first.

The second usual suspect is scripts: from plugins, chat boxes, social media feeds, several tracking codes, and fonts loaded from more than one place. Each was added for a reason, and none has been removed since.

Pick two fixes

Speed tests list many recommendations, and doing them all at once is a reliable way to break a site. Choose two.

The first is almost always the heaviest item: resize and compress the photographs, and replace a slideshow with one good image. On an image-heavy page, that is most of the weight gone.

The second is whatever holds up the first screen: scripts and fonts that must load before anything shows. Removing a plugin nobody uses, or loading the chat box after the page, brings the first screen forward.

Everything else on the list can wait for the next round.

A worked example

A gymnastics club with several hundred members has a website that parents use mainly for one thing: checking the schedule, often from the parking lot at pickup time.

The coaches thought the site was quick. They used it on the office computer, on a fast line, with the site long since stored in the browser.

A mobile test told a different story. The home page weighed over eight megabytes, mostly six competition photographs in a slideshow, straight from a camera. The screen stayed white while a calendar widget and two social feeds loaded.

The two fixes were chosen in an afternoon. The slideshow became one resized photograph, and the social feeds were replaced with plain links to the club's pages. The calendar stayed, set to load after everything else.

The schedule link now appears on the first screen before the photograph has finished arriving, which is what the parents were waiting for.

Test the pages people use

The home page is where most owners start testing, and it is not always where customers arrive. Search sends people straight to service pages, menus, schedules and contact pages.

Check analytics for the five pages where most visits begin, and test each. A menu saved as a three-megabyte image costs as much as a slow home page.

Then open each on your own phone with Wi-Fi switched off. The numbers describe the wait, and holding the phone shows how long it feels.

Test again after changes

Speed does not stay fixed. A new slideshow, a plugin, an embedded video or another tracking code each adds weight, and nobody runs a test before adding them.

Write down today's page weight and request count somewhere you will find them. Test again after any change to the main pages, and every few months regardless.

A number that has crept up is the signal to ask what was added.

The counter-case

Not every site needs this attention to the same degree. A wholesale supplier whose buyers order from office computers will find the mobile numbers matter less than a restaurant or a trade would.

Chasing a perfect score is not worth the cost either. Once the page is light and the first screen appears quickly, further gains are small compared with the effort.

For most local businesses, though, the customer is on a phone. Run the test, find the heaviest item, and fix two things.

The first round

  1. Run the home page through a mobile test.
  2. Note the weight and requests somewhere safe.
  3. Find the heaviest single item on the page.
  4. Resize and compress the photographs.
  5. Remove one script nobody would miss.
  6. Test the five busiest pages as well.
  7. Repeat the tests after every change.

Step three is what keeps the work small. Most pages have one item, usually a photograph or a slideshow, that outweighs everything else, and finding it turns a long report into a single task.

One rule the tests often flag, and which is usually a quick fix, is covered in compression switched off on the server.


Frequently asked questions

Why does the site feel quick when I check?

You test it at the office on a fast connection, with the site already stored in your browser. A customer on mobile data loading it for the first time waits much longer.

Which free tests should I use?

Google's Test My Site and PageSpeed Insights for an overview, and WebPageTest for a filmstrip showing what appears on a phone screen while the page loads.

What do page weight and requests mean?

Weight is the total size of everything the page downloads. Requests are the separate files it asks for. On mobile data, both add waiting time.

Should I aim for a perfect PageSpeed score?

No. The score counts rules followed, not the wait a customer feels. Use its list as a checklist and judge progress by page weight and loading time.

What usually makes a page heavy?

Photographs uploaded straight from a phone or camera, often in a slideshow. Scripts from plugins, chat boxes, social feeds and tracking codes usually come second.

Which two fixes matter most?

Usually resizing and compressing the photographs, then removing or delaying the scripts that hold up the first screen. Everything else can wait for a later round.

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.

Never tested your site on mobile data?

Send us the address. We will run the tests and tell you which two fixes to make first.

Start a Conversation