2026-08-07

Website carbon footprint: the hidden cost of your tags

You sell products with a low impact supply chain. You publish an annual sustainability report. Your packaging is compostable and you can name the farm.

Then someone opens your site in the developer tools and counts fourteen requests to companies you have never met.

That is not hypocrisy. Nobody sat down and decided to do it. A pixel got added for a campaign in 2023, a chat widget came with a support contract and a consent tool arrived to ask permission for the other two.

Still, if your brand promise is that you measure and reduce what you use, your website is the one place where a visitor can check that claim in thirty seconds.

This article is about what actually moves the number, how to measure your own and where the line is between a real claim and greenwashing.

Digital is physical

The mental model that causes the problem is the word cloud.

There is no cloud. There is a building with servers in it, a network moving bytes across the world and a phone in someone's hand doing work with a battery.

Every request your page makes runs that chain. A byte that never leaves the server costs nothing anywhere along it.

One tool that answers your questions beats five that each add weight. That is the idea behind a five minute weekly check.

"If it's free, the Earth pays the price."

Gerry McGovern, World Wide Waste, 2020

McGovern's book is about the scale of digital waste and that line lands hardest on exactly the tools we install without thinking. A free analytics tool is not free of cost. It is free of invoice.

Why the numbers you see are estimates

Before the practical part, a warning that matters more for your brand than for anyone else's.

Website carbon calculators produce a number in grams of CO2 per page view. Those tools are useful and the people behind them are careful. The models still involve significant assumptions: the energy intensity of the network, the mix of the electricity grid, how much of the load sits with the visitor's device and how often a returning visitor hits a cache.

Different models give different answers for the same page. Ranges of two or three times are normal.

That does not make the exercise pointless. It makes one use of it valid and another one risky.

Valid: comparing your own page before and after a change, with the same tool and the same method. The relative improvement is real even when the absolute value is uncertain.

Risky: printing an absolute figure on your homepage as if it were measured rather than modelled.

For an ESG focused brand that distinction is the whole ballgame. A competitor or a journalist who checks your method should find caution, not a rounded number with a leaf icon next to it.

Still with me? Good, because the ranking of what to fix is not what most people expect.

What actually moves the number

Weight and requests are what you control. Here they are in the order they usually matter.

• Images and video. Almost always the largest share of bytes. An unoptimised hero image can outweigh every script on the page combined. Modern formats and correct sizing beat everything else on this list.
• Video that autoplays. The single most expensive thing a page can do, because it keeps consuming while the visitor reads.
• Third party scripts. Individually small, collectively significant and worse than their file size suggests because each one adds a connection and keeps the device's processor busy.
• Fonts. Several weights of a custom font are easy to add and rarely all used.
• Caching and repeat visits. A returning visitor who downloads everything again is pure waste. This is a server configuration, not a design decision.

Notice that analytics is on the list but is not at the top. Anyone telling you that switching analytics transforms your site's footprint is selling you something and I would rather you heard that from me.

The honest version: your tags are a meaningful slice, they are the easiest slice to remove and unlike your images they also carry a privacy cost.

Measure your own site in fifteen minutes

Three steps and you can repeat them after every change.

1. Count your bytes. Open your page with an empty cache, press F12 and read the transferred total at the bottom of the Network tab. Note the number and the request count.

2. Split it by party. Sort by domain and separate your own domain from everything else. Now you know what share of your page belongs to companies you do not control.

3. Run a carbon calculator, twice. Once now, once after your changes. Record the tool and the date alongside the result, because the models get updated and you want to compare like with like.

Do this for your three most visited pages, not just the homepage. Traffic weights the impact and on most sites the homepage is not the most visited page.

What you can honestly claim

Say what you did, not what you saved.

Good: "We removed six third party scripts and cut our homepage weight by 38%." That is a fact you measured and can reproduce.

Also good: "We use analytics that sets no cookies and adds a few kilobytes to each page."

Risky: "Our website is carbon neutral." That needs an audited method behind it and it invites exactly the scrutiny that a modelled estimate cannot survive.

Bad: any absolute figure with no method, no date and no tool named next to it.

The reason to be strict here is not caution for its own sake. Your audience chose you because you are careful about claims. An overstated website claim damages that more than a heavy homepage ever would.

Where StatJolt fits in

StatJolt is a few kilobytes and loads asynchronously. It sends a small request per page view rather than a stream of them.

Because it sets no cookies, the consent tool that exists to ask about cookies can go too. That is usually the larger saving, since consent platforms are heavier than the analytics they gate.

The servers run inside the EU, in Germany and Finland, where grid intensity is generally lower than the global average.

I will not tell you this makes your site carbon neutral. It removes two scripts, some bytes and a layout shift and it does so without giving up the measurement. That is a claim you can put in a report and defend.

If you only read one box

Your page weight is physical and images and video dominate it so fix those before you touch anything else. Third party tags are the easiest weight to remove and the only kind that also carries a privacy cost. Measure your own bytes before and after with the same tool and publish what you changed rather than an absolute carbon figure you cannot defend.

Back