“Data Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is by Andy Atherton, SVP at AppNexus. It's the second in a series on the emergence of programmatic reserve.
A few weeks back, I laid out the case for a closer look at programmatic reserve. In particular, I focused on the need to better understand the business requirements behind today’s non-programmatic buying processes to deliver effective solutions in this area.
Some readers may have taken a moment to reflect on how odd it is that in 2013 – nearly 20 years into online advertising – we haven’t done a better job of automating a painfully inefficient process. Over the last five years we have created RTB from scratch and dramatically scaled this brand new and highly efficient technology. Meanwhile, for the rest of the digital display buy, we are still emailing RFPs and proposals, signing I/Os with pens and in some cases even still faxing them back. This is not just urban legend. I checked and it’s true. Some signed I/Os are still faxed.
So why has this "traditional" digital buying process been so resistant to change? I think it’s a combination of four main factors: a disproportionate focus by the tech community on lower-funnel solutions; agency incentives; agency organizational structure; and the chicken/egg nature of changes to established processes.
Clash of the Hemispheres
The traditional digital media planning and buying process is a complex combination of right-brain and left-brain. The Silicon Valley/Alley tech ecosystem that has been delivering solutions to the online advertising market is left-brain-dominated. Online ad technology solutions have driven impressive results in areas where metrics rule -- witness RTB display, or even search for that matter. But unfortunately, the approach that unlocked so much value in a direct response/lower-funnel setting has done a relatively poor job of improving media buying for campaigns with less direct or longer-cycle metrics.
Workflow improvements that accommodate the right-brain thinking that adds real value in traditional digital buying, often with an upper-funnel focus, simply have received less attention than technology for finer and finer slicing of audiences. However, research suggests that such technology has dubious value for upper-funnel campaigns. The traditional digital buying process has been too difficult for too long.
Follow the Money
According to 4As research, "Fees for service, based on agency labor, are the predominant form of compensation in the marketing services industry." In a world where agency profits scale with labor costs, there simply isn’t much incentive to be efficient with labor.
Over the last several years, agency holding companies have created centralized "trading desk" units to manage RTB buying. This structure reflects the learning curves and economies of scale inherent in the training and technology required to be effective in this arena. But it also offers agency management an opportunity to experiment with different economic models. These efforts have received more focus from management over the last several years than improvements to the traditional model. Technology vendors have happily pitched in, as this is fundamentally a left-brain effort.
Decentralization Is a Double-Edged Sword
Despite the growing role of the centralized trading desks, the organizational structure of the modern agency remains highly decentralized. Yes, holding companies dominate, but there is a very high degree of operational independence for individual agencies and even for individual client teams within most agencies. This structure is a reflection of clients’ needs; marketers are geographically distributed and have different cultures and goals. For example, American Express has different needs in different areas than Taco Bell, yet Digitas can serve them both well.
However, this structure also makes top-down, centralized change difficult to drive. Even if senior management was committed to improving the efficiency of the traditional digital media buying process, it would be difficult to drive ground-level adoption of the complex enterprise software necessary to accomplish this improvement.
Trading desks obviously provide a counter-example, but I would argue that their growth has been facilitated by the fact that these units focus primarily on lower-funnel budgets. This lower-funnel focus makes the differences between clients less important and media/technology more fungible across them. To extend our example, retargeting campaigns for American Express and Taco Bell have much more in common than their brand awareness campaigns.
For lower-funnel campaigns, then, the benefits of centralized trading desks can overcome the fundamentals that resulted in the decentralized agency structure in the first place. Not so for upper funnel campaigns, which make up a large chunk of agency digital budgets.
Markets rely on both supply and demand. It’s hard to get one without the other. This point is fairly straightforward, but that doesn’t diminish its importance. A player that had aggregated in one catalog substantially all the supply that traditional digital media buyers are used to accessing might be able to drive the buy-side to streamline the traditional digital buying process.
However, to build this catalog in the first place would require a ton of effort and cooperation from hundreds of publishers, which have a wide variety of different requirements, yield management strategies and sales channels amongst them. This effort and cooperation would be very difficult to achieve without lots of aggregated demand. Aggregating demand is difficult for the reasons we have discussed. And so on. For the sake of argument, we’ll call the sell-side the "chicken," but we could just as easily run the example in reverse.
Each of these four hurdles is significant independently. Collectively they are truly daunting, but the prize is great. The winner in this market will facilitate the flow of 11 figures in digital spend – a market even larger than RTB.