Desktop tests assume a faster machine and connection, so the same page scores better. Both are measured separately, and the one to prioritise is whichever matches most of your traffic.

Two datasets, one site

Speed is not measured once and reported two ways. Desktop and mobile are collected and reported as separate sets of data.

A page can pass comfortably on one and fail on the other, and both statements are true simultaneously about the same page.

Which means asking what your site scores is an incomplete question, and looking at only one of the two gives an incomplete answer.

Most businesses have only ever seen one number, usually whichever their testing tool defaulted to, and are surprised by the other.

Why desktop scores better

The first is the biggest factor and the least obvious. A page carrying a lot of script is doing work, and a phone processor takes considerably longer over the same work than a desktop one.

The gap is normal

Worth stating clearly, since a poor mobile score alongside a good desktop one reads as a fault.

A meaningful difference between the two is expected on almost every site, and a site scoring identically on both is unusual rather than exemplary.

The question is not whether there is a gap. It is whether the mobile figure passes on its own terms.

Comparing the two against each other produces no useful conclusion, because they are measuring different situations rather than the same situation twice.

What the comparison does tell you is where the weight is: a large gap points at scripts and processing rather than at file sizes, since a big image is slow on both.

Which one to prioritise

Decided by your own traffic rather than by any general rule.

Look at your device split in analytics, which most businesses have never checked and takes a minute.

For most local and consumer businesses, mobile is a large majority, frequently around seventy percent, which makes the mobile figure the one that matters.

For some business-to-business services, professional tools, and considered purchases researched during a working day, desktop is a much larger share and the balance shifts.

Fix the one your customers actually use, and treat the other as secondary rather than equal.

Where they are close to even, fix mobile first anyway, since a page fast on a phone is fast everywhere and the reverse does not hold.

Test on the right thing

The practical error that produces false reassurance.

Checking your own site on your own machine, on office wireless, tells you almost nothing about the mobile experience.

Testing tools let you choose the profile, and the mobile profile deliberately simulates a mid-range phone on a modest connection, which is why the number is worse.

That is not pessimism, it is a reasonable approximation of a real customer standing outside in a patchy signal area.

Where you want a genuine check, load the site on an actual phone, on mobile data, away from your premises, which remains the most honest test available.

A worked example

A business tested their homepage and reported a good result, which they had taken from a desktop test.

The mobile figure for the same page was substantially worse, and the field data showed it failing on how quickly the main content appeared.

Their analytics showed seventy-four percent of visits were mobile.

So the number they had been reassured by applied to a quarter of their visitors.

The cause was a set of scripts for a chat widget, a review carousel, and two tracking tools, which a desktop processor handled quickly and a phone did not.

Removing the carousel and deferring the chat widget improved the mobile figure considerably and changed the desktop one barely at all.

That asymmetry is the characteristic sign of a script problem rather than a file size one.

Do not average them

A small point that prevents a misleading habit.

Reporting a single blended figure hides exactly the information that matters, since a good desktop result will mask a failing mobile one.

Where somebody reports your site's speed as one number, ask which profile it came from.

The same applies to any dashboard or supplier report: two figures, labelled, or the report is not saying anything actionable.

This matters more now that both are used as signals, since a single blended number cannot tell you which of the two is at risk.

Field data lags behind

The expectation to hold when you make a change.

A test result changes the moment you fix something. The data collected from real visits is a rolling average over a period, so it moves gradually.

Which means a fix made today will show in a test immediately and in the real data over the following weeks.

Do not conclude a change failed because the field figures have not moved after a few days.

And do not make three more changes in the meantime, since you will not be able to tell which of them the eventual movement belongs to.

The counter-case

Both numbers matter less than they attract attention.

A site passing on both is done, and further improvement above the threshold buys nothing measurable.

The gap also becomes a distraction when a business starts optimising for a test rather than for a customer: stripping useful features to improve a score is a poor trade if the feature was producing enquiries.

And a genuinely fast site with weak content, no reviews, and a thin listing will still lose to a slower competitor who has those things.

Check both, fix mobile if it fails, and then go and work on something that produces customers.

The check

  1. Test the same page on both profiles.
  2. Expect a gap and do not treat it as a fault.
  3. Check your device split in analytics.
  4. Prioritise whichever most customers use.
  5. A large gap points at scripts, not file sizes.
  6. Load it on a real phone on mobile data.
  7. Wait weeks for field data to reflect a fix.

Step three is the one that makes the rest meaningful, and it is a number most businesses have never looked at.

How the measurements work is covered in measuring speed the way Google does.

Frequently asked questions

Why does my mobile score look so much worse?

Because desktop tests assume a faster processor, a better connection, and more memory. A page carrying a lot of script takes far longer on a phone processor.

Is a gap between them a problem?

No. A meaningful difference is expected on almost every site. The question is whether the mobile figure passes on its own terms, not how it compares to desktop.

Which one should I fix?

Whichever most of your customers use. Check your device split in analytics. For most local businesses mobile is around seventy percent.

What does a large gap tell me?

That the weight is in scripts and processing rather than file sizes, since a large image is slow on both. A script problem improves mobile sharply and desktop barely.

Should I report a single blended score?

No. It hides exactly the information that matters, since a good desktop result masks a failing mobile one. Ask which profile any single number came from.

Why has my field data not moved after a fix?

Because it is a rolling average over a period, so it lags by weeks while a test changes immediately. Do not make further changes while waiting.

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.

Only ever seen one speed score?

Test both profiles, then check your device split. Most businesses have been reassured by the number that applies to a quarter of their visitors.

Start a Conversation