There is a particular school of thought in ad tech that espouses the point of view that brands should refrain from “helicopter marketing” in the way they conduct their programmatic media activities. This line of thinking asserts that people should not micromanage or “hover” over the excellence of machines.
Wait a minute!
While machines are increasingly important and useful in achieving efficiency and efficacy in digital advertising, I firmly believe that marketers need to still hold human beings ultimately accountable for their programmatic media and marketing automation plans. Only humans can manifest and take advantage of serendipity. Only humans possess the nuance and instinct to do the particularly subjective risk-reward assessments that can lead to marketing brilliance. Robots are not set up to leverage risk and brilliance.
So how did we get to this point where so many marketers seem to be urged to rely more on artificial intelligence than on their human counterparts? It boils down to the fact that many technology companies have developed business models that are based on the arbitrage of media — buying media cheaply and then reselling it at a premium. This promotes black-box tactics that are in the best commercial interests of these players’ business models, thus leading to the elevation of opaque AI, which may not necessarily best serve clients’ marketing models.
This trend has been coupled with a propensity on the part of many marketers to place too much emphasis on hitting performance-based targets at the expense of truly understanding the dynamics and behavioral signals of their customer base.
Marketers should strive for a more informed and nuanced perspective on who in effect is being reached; which type of creative approach is most effective; and which publishers and channels are making the most impact. Marketers are doing a deal with the devil if they are outsourcing to third parties without obtaining new learnings in exchange.
There are three reasons why marketers need to manage AI transparently:
Seeing the Signals: While AI is important and relevant to achieve scale in an increasingly fragmented media market, transparency is paramount in optimizing spending and achieving goals. This lack of vendor and agency AI transparency results in the marketer unduly taking the risk in programmatic media.
In the evolution of media buying, risk swings on a pendulum between the buyer and seller. With opaque, black-box models, vendors are saying they will own the risk with no questions asked and no data shared. Consumer signals are irrelevant as long as conversions are made. The paradox is that this approach actually increases the risk for the brand.
Understanding the Attribution: Opacity does not belong in advertising models. So even performance marketers should demand insights and the “how”. Full transparency leading to the sharing of all signals should be standard operating procedure; they shouldn’t be hoarded by third parties as the “secret sauce.” Full transparency also helps marketers understand the true value of algorithmic solutions, and protects against impression-delivery timing and location games that might boost opaque AI results at the expense of other efforts. Opacity should not be exploited to keep marketers tied to vendors.
Protecting Proprietary Insights: Marketers may not realize that some opaque AI models essentially work to aggregate their signals from the noise across other customers. Marketers are thus participating in a great cookie collective for the third-party company. While some collective activity may end up as an exchange of data for the greater good of all members, certainly some data-rich marketers end up benefiting the data-poor marketers. Marketers should not let others hijack their data and mask how they are utilizing it.
Advertising always has been -- and will continue to be -- the marriage of art and science.
That goes for media buying as well. The pendulum can’t swing too far toward science. Because as smart and efficient as robots are and can become, art is still best employed by creatures driven as much by emotion and instinct, honed by insight — that’s right, human beings! For all of its impressive qualities and attributes, artificial intelligence still has difficulty keeping up with media innovations like the proliferation of new sites, apps and creative content.
So how do we effect a shift away from opacity and toward full transparency in digital marketing and the use of AI? How do we go about weaning ourselves off of this current addiction to black-box advertising? To me, the prescription is clear, if not challenging. I believe it is incumbent upon all client marketers to summon the courage and knowledge required to demand that they control the marketing activities, insights and use of their brand’s audience and advertising data. Only through transparent collaboration with agencies and vendors will marketers be ultimately performing due diligence and fulfilling their fiduciary duties to their brands.