The question people ask is how much CO₂ their website produces per page view. The answer is that nobody knows, and that a useful estimate is available anyway.
Nothing measures a website’s emissions directly. What exists are models: take the data your page transfers, apply assumptions about the energy required to move and process it and the carbon intensity of that electricity, and produce a figure.
Understanding your website carbon footprint means understanding that sentence, because it determines what you can honestly do with the number afterwards.
Website Carbon Footprint: How to Measure and Reduce It
What actually drives the figure
- Page weight, above everything else. This is measured directly, and it is the input every model depends on.
- Where the visitor is. Mobile networks are more energy-intensive per byte than fixed connections, and the carbon intensity of electricity varies enormously by country and by hour.
- Hosting. Whether the provider is treated as renewable-powered, and what that claim is taken to cover.
- Repeat versus first visits. A returning visitor with a warm cache transfers a fraction of what a first-time visitor does. Sites with high returning traffic look considerably better than their first-load figure suggests.
- Traffic volume. The multiplier that turns a per-page figure into something meaningful.
Getting a figure for your own site
Start with your five to ten most-visited pages, taken from analytics rather than assumption. The homepage is rarely representative.
Record page weight and request count from your browser’s network panel with the cache disabled, and note the conditions: device, connection, first or repeat visit, and the date.
Run a carbon calculator on the same pages. Note which tool and which version, because different models produce different results for identical pages and switching between them makes your trend meaningless.
Then multiply by monthly traffic. This step reorders your priority list more often than not, since your heaviest page is frequently not your busiest.
Agnikii Digital’s FREE Website Carbon Tool
What the number is good for
Comparison, mostly. Against your own earlier measurement, or between two versions of a page.
If your homepage transfers four megabytes and a competitor’s transfers six hundred kilobytes, no disagreement between models changes which is heavier. Relative statements hold where absolute ones wobble.
What it is not good for is precision. Anyone quoting a website’s emissions to three decimal places has misunderstood the input.
Reducing it
- Images. Display dimensions, modern formats, lazy loading below the fold, no autoplay video. Largest single win on almost every site and no design compromise.
- Third-party scripts. List every external domain and find an owner. Unclaimed ones are free to remove.
- Fonts. Two families at most, subset to the characters used.
- JavaScript. Most expensive per byte, most effort to fix, so last.
Re-measure after each change under the same conditions. 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.
Accessibility sits alongside
Clear navigation reduces how many pages somebody loads to find what they came for, and an avoided page load is transfer that never happened.
Semantic HTML reduces script dependence. Text alternatives keep content usable when media fails.
Work towards WCAG 2.1 Level AA on its own merits. The practices overlap substantially with efficiency without either causing the other.
Saying anything about the result
Environmental claims must be accurate and substantiated. The Competition and Markets Authority’s Green Claims Code sets the standard.
State the figure, the tool, the conditions, the date and the scope, and describe it as a modelled estimate covering website delivery. A specific measured reduction is defensible; a neutrality claim generally is not.
Our website sustainability audits produce that figure with the method documented.

