Carbon emissions, net zero media buying are the latest greenwashing shiny objects

Carbon emissions, net zero media buying are the latest greenwashing shiny objects

I've witnessed a disturbing amount of discussion, conference presentations, and even service offerings around reducing carbon emissions from programmatic media buying -- e.g. IPG Mediabrands to offer clients net-zero media buying.

It's entirely greenwashing BS. Why? Because all of this can be avoided entirely; it was entirely unnecessary anyway. You don't need to pay extra to solve problems you could have avoided entirely, in the first place. If an advertiser bought digital ads from a publisher, each ad impression should only require a a single call to an ad server. By buying ads programmatically, through real-time bidding (RTB), hundreds of ad tech companies insert themselves to "intermediate" the process of serving that single ad impression.

The PageXray below for CNN.com shows that even for such a mainstream site, the following large numbers of calls are observed for a single page: Adserver Requests: 581, Tracking Requests: 360, Other Requests: 99. "99.999% of the bandwidth and compute power used in RTB (real time bidding) does not result in the serving of an ad impression." This is because when a single ad opportunity comes up, bid requests are sent out to dozens of ad exchanges, which in turn present the opportunity to hundreds of advertisers; dozens of advertisers return bids, and only one wins the bid and the right to serve the ad into that ad slot. Multiply this by the number of ad slots on the page, the number of pageviews per publisher per day, the number of sites active in programmatic exchanges, etc. and you can start to imagine the enormity of the bandwidth and compute power consumed, which did not result in any ads being served. Using a simplistic 10% win rate means the other 90% didn't win and the bandwidth and compute did not result in an ad, but still generated carbon emissions. This is for legit sites.

No alt text provided for this image

If you take into account the other 90% of programmatic ad impressions that are manufactured by fake traffic on fake sites and fake usage of fake mobile apps, "100% of the carbon emissions generated by serving ads to bots are wasted because none of it produces a useful advertising outcome."

All of this recent chatter about the carbon footprint of digital ads is just the latest load of crap that ad tech vendors and agencies are trying to get advertisers to pay for. Supply-path optimization (SPO) came before it and is still being packaged up for sale and publicity. "Horizon Media and Magnite ink SPO deal" reads a recent headline. Reporter Ryan Barwick astutely asked "Why should advertisers have to pay more for SPO?" The messy and leaky supply paths can be avoided entirely by NOT buying from 25 - 35 exchanges at a time. The table on the left (below) shows a typical programmatic campaign where nearly 30 exchanges were left turned on. If you look closely 98% of the ads came through the top 5 exchanges, and 78% came from the top ONE exchange. So if you turned off (by unchecking) the other 25 exchanges, and just left those top 5, you would have solved the supply path optimization crap yourself. No cost. Took 3 minutes (depending on how quickly you can uncheck checkboxes).

No alt text provided for this image

Most exchanges are selling the same inventory anyway and the vast majority of bid requests are duplicates for the same ad opportunity. In fact there's something called "win rate laundering" which you may not be aware of yet. Why do win rates still appear to be reasonable -- e.g. 10% (1 in 10) -- if there were so many thousands of duplicated bid requests for the same ad opportunity? Your average win rate does not de-dupe all of the identical bid requests sent to a dozen different exchanges. It's the same ad opportunity/bid request but the reporting doesn't know it's the same. So you end up unknowingly bidding against yourself and paying more for the ad. That's exactly what sites and supply-side platforms want to happen so they can make more money. All of these forms of fraud, call it "shenanigans" if you want, are conveniently hidden from view by the processes of RTB.

You can avoid ALL of the above by buying direct from a handful of good publishers. So be more concrete: 1) create a list of domains and apps for an inclusion list, 2) identify the least number of exchanges possible, like 2 - 5, to include in the campaign, and 3) there is no step 3. You're done. You've cut out all the unnecessary supply paths; you've drastically reduced the bandwidth and compute used to serve ads (and the carbon emissions); and you've avoided most of the ads going to fraudulent sites and apps.

By doing this yourself, you can avoid all the greenwashing shiny objects like "net zero media buying" and "supply path optimization" that ad tech vendors and agencies are trying to get you to pay for to solve problems that could be entirely avoided in the first place.

I'll leave you with a final PageXray of dailymail.co.uk Adserver Requests: 2,161, Tracking Requests: 1,354, Other Requests: 301

No alt text provided for this image
https://pagexray.fouanalytics.com/q/dailymail.co.uk?f=
Abdul Mazed

Freelance Writer @ Self-employed | Writing and Reviewing

2 年

Great article

Niki Banerjee

Principal Solutions Consultant at Scope3

2 年

Great explanations Dr. Augustine Fou. As Keaton suggested, we'd love to take you through our model and get your feedback.

Slavina Dimitrova

Marketing & Digital Lead | Mentor | Community Building | System Thinking | Doughnut Economics Advocate | Leading with Empathy | Global | Local | B2B | B2C | Non-profits | Businesses

2 年

Thanks for the insightful article Dr. Augustine Fou

Keaton G.

Transformational Digital Marketing Leader | Future-Proof Strategies

2 年

I would like to add some clarity here. You are correct in that whitelisting premium publishers will lower emissions dramatically. There are far less ad calls on publishers of legitimate content than the long tail. Up to 10x lower emissions. A typical campaign for a whitelisted buy might generate 55kg of carbon, whereas the long tail would be 550kg. All to do with streamlined vendor relationships rather than opening the floodgates which is what many dodgy smaller sites do. I just want to address what you say is greenwashing though. Greenwashing is where a company claims to be green when it's not. Clear examples are paper bottles that have plastic linings, biodegradable bags that don't degrade etc. A second thing that could be greenwashing is using completely bad/misleading data. I believe that with detailed supply path data, the picture of emissions due to the nature of the internet's accountability vs say, calculating emissions from analogue things like a sports event is night and day in accuracy. So where does this leave what the big consortiums are talking about? They use accurate supply path data from scope3.com and I challenge you to get in touch to go through it, their methodology is on GitHub for all to see.

要查看或添加评论,请登录

社区洞察

其他会员也浏览了