Maybe Media Transparency is the Devil?
“The greatest trick the devil ever pulled was convincing the world he didn’t exist”
I first recall the famous line above as being recited by the character Keyser Soze in the movie The Usual Suspects, though there are several accounts of similar quotes much earlier in time. In light of recent reports regarding fraud, deception and waste in the world of digital media, I think it’s transparency itself the industry would like you to believe doesn’t exist.
Some History
We have been having this conversation for well over 10 years, and yet here we are in 2023 talking about the BILLIONS of ad dollars being wasted on an annual basis. Programmatic, or biddable media buying, began to take hold within large brands somewhere around 2010 and immediately stories of vastly improved ROI made waves. During my time at The Kellogg Company we were right there leading the charge. In fact, multiple times we publicly presented outcomes from some of the earliest investments in programmatic showing ROI gains of 5 to 6 times previous results.
It seemed as if we were on the cusp of a revolution in marketing effectiveness. Technology, data, and the “long tail” of the web had disinter-mediated any advantage of asymmetry large publishers previously held over ad buyers. But then we forgot, for every action there tends to be an equal and opposite reaction. As buyers learned to be ever more efficient in what they assumed was successfully reaching their desired audience, sellers (and anyone even remotely involved in supporting an ad transaction) responded in every way possible to game the system. Fast forward a few years and the reality is those ROI charts above did not continue to show improvements. Programmatic had allowed brands to drop the cost of advertising, but improved effectiveness just wasn’t there.
Why wasn’t effectiveness improving? Frankly, for all the reasons trumpeted in the headlines above. By 2014 we at Kellogg’s were on a crusade to holistically measure the viewability of ads, the amount of ads served to “non-humans” and percent of video impressions reaching completion with the audio on. Gaining this level of transparency into our ad buys was a monumental and frustrating endeavor, one that ultimately I say we, and all ad buyers, have thus far lost. For proof, just this summer the ANA released a first look from their most recent Programmatic Media Transparency Study, which included the following finding: “The study’s primary conclusion is that the $88 billion open web programmatic media ecosystem is riddled with as much as $20 billion in waste, a figure that represents 23 percent of the total open web programmatic media investment by marketers”.
So why is transparency into digital campaign delivery and performance so elusive? There are two collaborating drivers behind this challenge:
- Misaligned incentives
- Intentional and unnecessary complexity and confusion
Incentive Alignment
By now you have likely read much about AI and its potential implications on both the Marketing industry and in fact humanity at large. Because AI holds the potential to wield such immense power, the prevailing debate amongst experts (and frothy X users) is around the concept of “alignment”. We all have a good sense for the meaning of the word alignment in general, but I asked ChatGPT to explain specifically the concept of AI alignment. Here was its response:
AI alignment refers to the critical goal of ensuring that artificial intelligence systems’ objectives and behaviors align with human values and goals to prevent potential harmful consequences and ensure beneficial outcomes for humanity. It involves designing AI systems that reliably and ethically pursue the intentions of their creators while considering possible long-term impacts.
That is one meaty debate topic and not something I can necessarily advance through this blog. However, I will confidently claim that there is complete misalignment of incentives in digital advertising. Consider, what is an advertiser’s incentive to spend marketing dollars? To sell more of their products or services, naturally. Contrast that with the incentive of a publisher to put ads within their content. Their incentive is based on volume – of not just visitors, but specifically the combination of visitors and advertisements.
More impressions means more ad $ coming to the publisher. With that as the base incentive, one can begin gaming the individual components in the equation. For example, one entity may ask, does a visitor have to be human? One could make an infinite number of visitors if desired (we call these bots). Or, does the visitor actually have to see the advertisement? How about, how many advertisements on one page is too many? When my incentive is all about volume, it becomes really tempting to dial up volume any way one can.
But from an ad buyer’s perspective it gets much worse. The publisher is only the last point in what has become a very long supply chain of entities involved in delivering the advertisement. And guess what, every participant in that supply chain, from buying agencies, to ad servers to DSPs to exchanges to SSPs to verification companies to third party data providers and beyond is also typically INCENTIVIZED BY VOLUME. There are of course various approaches to attempting to better align incentives, but nothing very straightforward nor thus far successful. While we are all worried about AI alignment, maybe we can work on media incentive alignment at the same time. If not, it may be possible to do something about the second driver of the transparency challenge.
Complexity and Confusion
Ask 10 individuals who have spent years in marketing and AdTech to describe all the steps involved in making an ad for Pampers show up on any random website and you will get 10 different answers. That’s not because they all disagree, it’s just a reflection of how complex this system has become. Sadly, many involved in that system would say – that’s a feature, not a bug!
Like many messes that cost billions of dollars, this one began with the best of intentions. The rise of various technologies and data applications truly has created an opportunity for improved advertising performance along with enhanced consumer experiences. But then those pesky misaligned incentives got in the way and soon it was beneficial to have complexity and confusion which keeps ad buyers from knowing where exactly their media investment goes and how well it works.
You can draw a straight line from the growth of walled gardens, particularly Google and Meta, to the lack of transparency ad buyers receive today. As each of those properties gained more users and more ad dollars, their asymmetrical advantage allowed them to essentially tell advertisers “no” when asked for transparent measurement. And they did this with zero significant impact on ad spend from buyers. As the rest of the media ecosystem watched this play out, it became easier to put up barriers (real or perceived) to transparency. You want log level reporting for your programmatic buy? Oh we don’t allow that level of data granularity to leave our platform. You want third party verification for viewability and invalid traffic? Ok sure, (*whispers – but it will be based on data we feed them through an API, not their independent measurement*).
So here we are, advertisers being told true transparency doesn’t, or can’t, exist. And it is an easy answer to accept because the ecosystem is so complex that everyone feels confused, not even sure what questions to ask. How do you convince the world the devil actually does exist?
Demand Transparency
Marketers – you may not be able to perfectly align incentives, but you can protect your own. Your incentive is to make every $ spent work as hard as possible and you should demand the transparency required to measure performance. If a media partner will not allow you to see the data necessary to measure metrics important to you, perhaps spend elsewhere. Sure that’s easy to say, but I recognize it’s often harder to do.
To simplify getting to transparency, a good place to start is with a thorough assessment of your current digital media supply chain. You may be familiar with the ever expanding list of “Lumascapes”. I would challenge you – do you know what tools and partners your brands’ version of this includes? Map your digital media supply chain and use this to inform other questions you need answered. How many companies receive a slice of your media investment and how do each one’s incentives support those of your own? Who is willing to provide granular data, through privacy preserving means, that inform you as to what is working or not toward your objectives? Is every component of your media supply chain necessary and/or adding value?
These are simply the first of many questions you should ask during an assessment of your digital media supply chain. You will be surprised by many things you learn and way more empowered simply through the knowledge gained. Do this exercise yourself or ask a firm like us to help you, but either way DO NOT DO NOTHING. Transparent Partners has run such assessments for enterprise brands over the past 10 years, we have plenty of advice for any who are interested. But again, my concern is that you start asking these questions and taking action on behalf of your own media investment. And by all means, don’t let your media, technology and data partners convince you transparency doesn’t exist!
Aaron Fetters, Managing Partner, CEO
