Two people measure the same page. One reports 1.2MB, the other 3.8MB. Neither has made a mistake.
They measured different things, and until you know which one you are quoting, every number in this subject is unstable. That is the first and most useful thing to understand about website page weight, and it is skipped in most guides.
Website Page Weight: How to Measure and Reduce Bloat
Transfer size and resource size
- Transfer size is what actually crossed the network, after compression. This is what your visitor’s data allowance pays for and what carbon estimates are calculated from.
- Resource size is the uncompressed total after decompression. This is what the browser holds in memory and processes.
Text-based assets compress well, so the two figures can differ substantially. Images and video are already compressed, so for them the numbers are close.
Both matter. Transfer size governs network cost and carbon estimates. Resource size governs the work the device does. Report whichever you like, but say which one it is.
Measuring it consistently
The browser you are reading this in has everything you need.
Open the network panel, disable the cache, and reload your busiest page. The status bar at the bottom shows requests, transferred and resources. Sort by size to see what dominates.
Then record the conditions, because without them the number means nothing later: which device, which connection speed, first or repeat visit, and the date.
Repeat on your five to ten most-visited pages, taken from analytics rather than assumption. The homepage is rarely the heaviest.
Testing like a real visitor
Your office broadband and desktop processor hide problems your customers experience.
Use the throttling options in the network panel to simulate a slower connection, and test on an actual mid-range phone where you can. A page that feels fine on a laptop can be genuinely unusable on a four-year-old handset on patchy mobile data.
First visits matter more than repeat ones for most sites, since a large share of visitors arrive with nothing cached.
What counts as heavy
Be careful with universal targets. What is acceptable depends on your audience, your sector and what the page is for. A product gallery legitimately weighs more than a contact page.
The HTTP Archive page weight report tracks what the median page across the web transfers, and comparing against that is more honest than any single number quoted as a standard.
More useful still is your own trend. Measuring the same pages under the same conditions over time tells you whether you are improving, and controls for everything specific to your situation.
Turning measurement into a budget
A measurement you do nothing with is a statistic. A budget is a constraint.
Set a ceiling per page type: a homepage figure, a lower one for content pages, a higher one where imagery is genuinely the content. Base it on what you measured rather than an aspiration.
Then enforce it. Performance budgets configured in your build process fail a deployment when the ceiling is exceeded, which works considerably better than somebody remembering to check.
Without enforcement, an audit buys you roughly a year, because sites regain weight through ordinary content work rather than through anybody’s decision.
What the reduction gives you
A faster site, immediately and reliably. Better Core Web Vitals, which contribute to how page experience is assessed without being a decisive ranking factor.
Where performance and environmental impact both apply, treat them as the same lever: reducing unnecessary data transfer and processing supports a faster experience while lowering estimated digital carbon, though environmental figures remain modelled rather than directly measured.
If you publish a figure, state which measurement you used, under what conditions and on which pages. Transfer size and resource size are not interchangeable, and quoting the flattering one without saying so is the kind of thing that gets noticed.
Our website sustainability audits record both figures, with the conditions documented.