This FAQ is designed for content editors to understand how carbon emissions of websites are calculated in Imperial’s context.

How are my website’s emissions actually calculated?

We are using the Sustainable Web Design Model (SWDM) Version 4 (V4) as the standardised computation methodology to calculate website emissions using the main equation as follows:

C (gCO2e/GB) = D (GB) x E (kWh) x I (gCO2e/kWh)

Where

●       C = Carbon emissions

●       D = Data transfer, which uses page weight as a proxy

●       E = Energy intensity

●       I = Grid carbon intensity

 
The SWDM V4 uses page weight–the total size of the files downloaded when a page loads–as a proxy for data transfer (D). The grid carbon intensity (I) used is the global average grid intensity of 494 gCO2e/kWh based on global energy think tank EMBER’s data. The operational energy intensity used are 0.055 kWh/GB for data centres, 0.059 kWh/GB for networks, and 0.080 kWh/GB for user devices, based on data from IEA and ITA.

What is the Sustainable Web Design (SWD) Model v4?

The Sustainable Web Design Model is a "bottom-up" calculation method used to estimate the carbon footprint of a website. Version 4 (v4) is the current industry standard, providing a more granular and accurate view of energy consumption across the entire digital lifecycle. In short, it estimates emissions using the following equation:

Estimated emissions (gCO2e/GB) = Operational emissions + Embodied emissions

  • Operational: The emissions attributed to the use of the devices in a segment.
  • Embodied: The emissions attributed to the production of the devices in a segment.

It uses broad system boundaries to represent a comprehensive GHG footprint, segmenting the digital system into three key segments based on the allocated proportions in Table 1 below. 

  • Data centres: Energy required to house and serve data.
  • Networks: An allocation of energy used by networks to transfer data.
  • User devices: Energy used by end users interacting with a product or service.

 

Table 1: Breakdown of allocated energy for each system segment (‘proportion of total’), and internal breakdown within each segment (‘operational’ and ‘embodied’)
Breakdown of allocated energy for each system segment (‘proportion of total’), and internal breakdown within each segment (‘operational’ and ‘embodied’)

 

Data centers

Networks

User devices

Proportion of total

22%

24%

54%

Operational

82%

82%

49%

Embodied

18%

18%

51%

In calculating estimated emissions of a digital activity, the SWDM uses data from the International Energy Agency (IEA), International Telecommunication Union (ITU), global energy think tank EMBER, and other previous research. For a deeper dive, refer to the References section.

What other computation methodologies are there and how does SWD V4 differ from them?

Apart from SWDM, we are only aware of the OneByte Model as the other commonly used computation methodology in mainstream. The table below shows some of the main differences:

Methodology

Approach

Primary Focus

Best Use Case

SWDM v4

Bottom-up

Broader system boundary (as explained in FAQ 2)

General website benchmarking and content auditing.

OneByte Model

Top-down

Narrower system boundary–data centre and network only

Simple, high-level estimates for massive data flows (e.g., streaming).

If you are aware of other computation methodologies that may be more accurate or fit for purpose, please get in touch with the Digital and Sustainability Team to discuss.

Why was the SWD Model V4 chosen?

We use the Sustainable Web Design Model V4 because it offers the best balance of comprehensiveness and actionability:

  • Holistic View: Unlike older models that only looked at hosting, SWD V4 accounts for the user's device and the manufacturing of hardware. The broad system boundaries used by the model also allows application to a wide variety of digital activities, not limited to websites, but also apps, web browsing, file downloads, video and audio streaming, and API responses. (MightyBytes & Wholegrain Digital, 2025)
  • Granular and Bottom-Up: It allows us to see that over half of our impact happens on the user's end. This empowers content editors to reduce emissions by optimising images and removing heavy scripts.
  • Standardisation: It is the default model used by major tools like CO2.js, Ecograder, and the Website Carbon Calculator, ensuring our data is comparable to industry benchmarks. It is designed to work with existing well known standards like the GHG Protocol Corporate Standards. (MightyBytes & Wholegrain Digital, 2025)
  • Updated Data: V4 incorporates the latest "Ember" electricity grid data, reflecting the global shift toward renewable energy more accurately than older versions.

As a content editor, how can I lower the carbon emissions of my website?

The newly launched web sustainability guidelines are meant to help content editors think about how to lower the carbon emissions of their website, and contain practical and actionable recommendations on how content editors can design their content and information architecture to improve sustainability. 

In short, the primary goal is to improve efficiency–helping your audience achieve their goals in visiting your website as quickly as possible. One practical tip is to reduce page weight, which the SWDM uses as the proxy for energy use, by considering the common heavier web elements:

  • Delete Outdated Content: Regularly audit and remove old, unused pages and heavy assets that no longer serve a purpose.
  • Content Strategy: Ensure users find what they need quickly. Less time spent searching means less energy consumed by the user's device.
  • Optimise Images: Use lighter formats like JPG rather PNG.
  • Use Thoughtful and Purposeful Media: Avoid auto-playing videos or massive "hero" background animations.

What are the limitations of using the SWDM to calculate a website’s carbon emissions? 

The Sustainable Website Design Model (SWDM) offers an estimate of energy use, using data transfer as a proxy. Like all models, it has limitations, relying on best estimates for several variables where granular data is unavailable or cannot be collected. For example, the SWDM only fully accounts for front-end (client-side) performance, without comprehensive consideration of the backend (server-side) activity.

Consequently, while the SWDM can provide a rough indication of potential sustainability improvements for a website, content editors should avoid targeting specific numerical scores or assuming that good SWDM performance equates to a perfectly sustainable website.

References

Green Web Foundation (2026) CO2.js: Models. 2026. Thegreenwebfoundation.org. https://developers.thegreenwebfoundation.org/co2js/models/overview/ [Accessed: 30 March 2026].

MightyBytes & Wholegrain Digital (2025) Estimating Digital Emissions. July 2025. Sustainable Web Design. https://sustainablewebdesign.org/estimating-digital-emissions/.