Our host procures hourly-matched renewable energy, operates in a low-carbon grid region, and runs extended server lifecycles. No provider leads on everything, we chose deliberately and can explain the tradeoff.

The 30% came from a model that treats all bytes equally. Reduce bytes, reduce carbon, linearly. But network energy doesn't scale like that. Sending less data is genuinely better (faster pages, less device time, lower CDN cost) but the carbon number is an artefact of the model, not a measurement of what changed.

All media assets are compressed, correctly sized, and served in modern formats - following W3C WSG guidelines 3.2, 3.3, 4.3. This reduces page weight, improves performance, and lowers energy demand on end-user devices. We don't claim a specific carbon saving because the models aren't mature enough to support one.

There is no regulated methodology, no standardised threshold. That doesn't mean the rating isn't useful - it is. But a single page, one model, one moment tells you nothing about your website. Extrapolating a homepage rating to thousands of pages with different weights, functions, and traffic patterns is statistically indefensible.

Sampled across page types, weighted by real traffic - the average rating a visitor actually encounters on our site. We name the tool, acknowledge its limits, and state our sample size.

One page's per-visit estimate × total annual visits. That assumes every page has the same footprint, every visit is identical, and the model's uncertainty doesn't compound. A text article and a checkout flow don't weigh the same, your total could be off by an order of magnitude.

Same claim. Defensible evidence:

Built and maintained to the latest web sustainability guidelines, and measured to recognised standards, across the whole lifecycle from end-user to data centre. The guidelines exist. The standards exist. The organisations that start engaging today are the ones that won't scramble when this becomes the baseline.