Is the U.K. model workable here? Digital Campaigns and their Carbon Costs

By Aaron Yeardley

It is now essential for companies to quantify, understand, and reduce their carbon footprint. Yet, most companies only consider embodied carbon in their products and operational carbon from manufacturing and office emissions. But now, the paradigm is beginning to shift so that a digital carbon footprint is also required.

The UK’s advertising industry has activated the AD Net Zero Action Plan in a bid to reduce carbon emissions from digital advertising to net zero by 2030. The five-step action plan includes a “simple” list of actions that each business should follow to become real net zero. However, Action 3 is far from simple.

With the aim to reduce the carbon emissions from digital advertising, Action 3 aims to curb the emissions sourced from media planning and buying. This should enable businesses to make more informed decisions when choosing media agencies. But the caveat; Action 3 is totally reliant on a tool for the future called DIMPACT, which is currently in its prototype stage.

DIMPACT has begun making huge strides by enabling participating media companies the ability to calculate the greenhouse gas emissions associated with serving media content, including that from the ad delivery value chain.

However, for players in the media and advertising industries, quantifying carbon emissions any later than now is too late. Together, we all need to take an active role in enabling carbon reductions. We simply cannot wait for tools to do the job for us.

Calculating the environmental impact of digital advertising is no simple task. The various routes through the programmatic supply chain complicate calculations but the inherent inefficiency of digital advertising causes all routes to be environmentally damaging.

Here, we will depict the methodology involved in calculating the greenhouse gas emissions from one route through the programmatic supply chain. We aim to show that whilst the nature of digital advertising is complicated, the carbon calculations are possible.

The case study reported here considers header bidding services through the private marketplace. This is a popular route as it is where advertising is bought and sold automatically between exclusive users. This automation of ad buying and selling enables targeted advertising through data transmission, storage, replication, and processing. However, the sophisticated ecosystem uses powerful artificial intelligence algorithms to evaluate users based on their behaviour, demographic data, cookie data and other criteria to ensure each ad is shown to a targeted audience.

Thus, digital advertising uses a lot of power due to the amount of data being collected from users visiting digital platforms, the transmission of such data, and third-party replication of data. A typical scenario includes an end user visiting a webpage, agreeing to cookies, and opening a mobile radio that connects the end-user with dozens of data centres. Then trackers are used to track the users’ behaviours and deliver online ads. Finally, third parties obtain the valuable information in a practise referred to as cookie-syncing. All the programmatic supply chain must be accounted for in the carbon emission calculations.

In short, online advertising increases carbon emissions by increasing energy consumption end-to-end from four factors: 1) the amount of downloaded data increases, 2) the varying inter-transfer interval between the network, data centre, and end-user device resources, 3) the time required to access the payload content or application increases, and 4) the amount of active connections increases.

Thus, calculating the carbon emissions from one advertising campaign is challenging, but by considering each factor in the programmatic supply chain, it is possible. First of all, the carbon emissions sourced from one advertisements’ storage size can be easily calculated:

Then each of the four energy consumption factors must be included in the calculations. For example, each impression draws down the asset increasing the amount of downloaded data, so number of impressions can be factored into the calculation. Additionally, the number of third parties that store data from the advertisement can be scaled to give a value that one typical advertising campaign emits once it is placed on a website.

But, within the programmatic supply chain, a lot of energy has already been consumed by artificial intelligence algorithms collecting end user data and selling the advertisement to its targeted audience.

Similarly, to the carbon emissions produced by one advertisements storage size, the emissions from cookie syncing can be calculated by measuring the size of the data being collected from the end-user.

But the constant summation does not stop there, the data is constantly being stored, replicated, and processed at data centres. Now this is what leads to the ultimate question; to calculate the carbon emissions from digital advertising, what is the system boundary?

We could look at the whole picture, where emissions are sourced from:

• The size of an advertisement and how many times it is seen.
• The transfer of user data enabling smart targeted advertising.
• The power usage from a data centre storing the data.
• The power being used by an end-user viewing the advertisement.
• The embodied carbon within the data centre, servers, end-user devices, etc.

Clearly, deciding on a system boundary is important for carbon calculations. And the appropriate scope is dependent on the question being asked. If asking what your total carbon footprint is then all things must be considered. But when trying to reduce your carbon footprint, you may just be comparing one digital advertising media to another. Therefore, the choice of a system boundary must be well thought out and justifiable.

Aaron Yeardley is the Carbon Reduction Engineer for Tunley Engineering.

Glossary

• Carbon footprint is the total greenhouse gas (GHG) emissions caused by an individual, event, organization, service, place or product, expressed as carbon dioxide equivalent.
• Embodied carbon is the total GHG emissions generated to produce a product; It includes those from extraction, manufacture, processing, transportation and assembly in every component.
• Operational carbon is the total GHG emissions generated from business operations. It includes those from fuels and electricity used for energy and any indirect carbon emissions sourced from products used during business operations.
• Digital carbon is the total GHG emissions generated from digital devices.
• Net zero carbon is the sum affect of combining actions to reduce GHG emissions with actions to off-set them.
• Real net zero is when the greenhouse gas emissions through activities will equal zero with carbon removal activities included.