Tackling climate change isn’t just about building wind farms and buying electric cars. Every sector must do its part to reduce CO2 emissions so that we can avoid the worst effects of global warming; and we’re not on track. One of the main barriers to decarbonisation is financial. Cutting CO2 is often seen as a cost, with short term priorities getting in the way of long term benefits.
But there’s one area — thanks to the power of AI — where decarbonisation can be delivered without compromise, where every gram of CO2 saved also saves money. You may have never heard of it, but you’ve definitely encountered it: programmatic advertising.
The invisible supply chain with a pollution problem
Programmatic advertising is something few people think about, but everyone experiences. Whenever you see an ad in a digital setting — be it social media, YouTube, or a banner on a website — an advertiser has bid for the opportunity to get your attention. With trillions of ad slots available every day, this process has to be highly automated, or programmed, in order to function at scale; hence, “programmatic advertising”.
The problem is, advertiser budgets aren’t programmatic advertising’s only cost. Every ad served comes with a little puff of CO2, tiny in isolation, but massive when added up, 24/7, across the entire internet. Over a year, the invisible processes that keep digital advertising running produce 7.2 million tons of CO2, equal to the average annual emissions of 1.4 million US homes. That’s a big price tag for the planet just to deliver ads.
It gets worse. Not only does programmatic advertising pollute, it doesn’t do a particularly good job at what it’s supposed to do either. The Association of National Advertisers found that only 36 cents from every dollar advertisers spend programmatically reaches the other end of the supply chain. The rest is eaten up by fees from intermediaries, low quality placements (think websites full of clickbait and spam) or simply disappears in the notoriously murky web of technologies that make up the supply chain.
Big data? No problem
Both polluting and inefficient, programmatic advertising is the perfect candidate for some much-needed spring cleaning. But the very complexity that causes its issues has prevented meaningful change towards a leaner and more optimised supply chain. Just look at this diagram of the digital advertising ecosystem — how would anyone even know where to start?
Cleaning up programmatic advertising was a task beyond humans, but is exactly the kind of messy big data problem that AI excels at. What looks like an incomprehensible tangle of overlapping data to us is a simple matter of pattern recognition for machine learning algorithms.
By analysing the historic performance data of programmatic advertising campaigns, AI algorithms can identify which supply paths result in low performing, high emission outcomes and apply this training to optimise future campaigns. For advertisers, it’s a simple matter of plugging this AI-powered optimisation API into their platform of choice, then sitting back and letting the algorithm do the work.
With AI, we are able to account for a wide array of factors that can affect programmatic performance and emissions. The day of the week, the time of day, the device the ad is displaying on, whether the user is on Wi-Fi or mobile data, and the share of renewable electricity in the local energy grid all impact the CO2 emissions of an impression. There’s no limit to how many plates an AI algorithm can spin.
We can even account for unpredictable and unwanted downstream effects of optimisation. For example, there’s a phenomenon called the “rebound effect”, where reducing the cost of delivering ads results in more campaign activity. In short, if something’s cheaper, you buy more of it, and drive up emissions in the process. AI can account for this effect and ensure that emissions remain under target.
Now advertisers can’t afford not to decarbonise
The best part of optimising programmatic advertising for CO2 reduction is that it also results in campaigns that cost less and deliver more. It’s a triple-win for advertisers, switching the narrative on decarbonisation from a nice-to-have to a business imperative. While it would be fantastic if everyone reduced emissions for noble reasons, the fact is that nothing moves the needle like money.
At Greenbids, we’ve been moving the needle in campaign after campaign with our AI-powered optimisation engine:
● French electricity provider EDF decreased its cost-per-conversion by 17% while reducing campaign emissions by 13% ● Spotify reduced cost per view by up to 40% (depending on format) for a global YouTube campaign, while cutting carbon intensity by 38% ● Energy supplier Electric Ireland enjoyed 22% lower costs, a 31% performance boost, and a 67% CO2 reduction
Results like these simply would not have been achievable without AI. Most importantly, AI makes programmatic decarbonisation scalable, which is crucial in driving meaningful change and making a positive impact at a global level. Cost-saving CO2 reduction can be achieved in any programmatic campaign, across any media, anywhere in the world, in a few clicks.
