Where does the money go when you trade advertising? Who gets paid to help buyers and sellers make the deal? To frame the proceeding, we will be discussing the economics associated with RFP based buying, not RTB. (RTB is significantly more expensive to transact)
Let's start at the highest level, the publisher and the advertiser. Each side has a team of people in-house, an agency, or other service provider to manage their side of the advertising process. To enable these teams to work efficiently, a myriad of technologies have been created. To emphasize the point, here is a short list of functional technologies and people needed to make things happen.
Technologies: Ad serving, billing and reconciliation, data management platforms, inventory management systems, tag management systems, supply side platforms, demand side platforms, ad exchanges, private marketplaces, analytics, media planning tools, yield optimization, re-targeting, creative optimization, and verification.
People: VPs, digital media directors, digital media supervisors, media planners, assistant planners, media buyers, campaign supervisors, campaign managers, ad traffickers, campaign coordinators, traffic managers, client services managers, analysts, and directors of ad operations.
Seriously? I'm out of breath just reading that. The first time that dawned on me, I had a hard time comprehending the complexity of what it actually takes. In truth, you have to respect that. Anything so intricate must take highly trained professionals to enable. So what does that mean in economic terms? About 10% goes for making a campaign happen on the buy side, 10% for making the campaign happen on the sell side, and 10% for the technology providers that sit all over the supply chain. In short, its expensive, very expensive.
So why have machines not replaced much of the tedious work done by people in advertising like it has been in so many other industries? Some will say that it is a level of complexity machines can't handle. Some will say that agency and publisher business models have rationalized their costs based on head count. Still others will say that advertising is a business of relationships. That's all true, but machines use rules. Machines don't care how many rules there are or how complex they are. As long as the rules people use can be defined in a way machines can understand, tools can be created.
As much as we all want to believe that we live in a world swimming in marketing and advertising data, if it can not be used to facilitate automation or to dramatically improve manual media buying and selling processes, what good is it really? In fact, if people are so much better, why do they need data? Simple, people use rules and data just like machines. The difference is that they are much better at making abstract judgement calls that machines completely fail at.
So, what are the economics of advertising and how are they changing? First lets look at what is being bought and sold - space in a publication that if left unused has no value. Next, the different kinds of things there are to buy and sell - pretty much an infinite combination of media types, unit sizes, context, audience attributes, etc. The economics of advertising transactions are about people solving complex pricing and selection problems and getting paid to do it.
Economics teaches us that real world situations present numerous examples of transaction costs, each of which can be grouped into three broad categories: geography, technology/infrastructure, and institutional/policy related transaction costs. Geography, otherwise referred to as the transaction costs of transportation do not impact advertising transactions in the modern age. Technology has transaction costs too, but its impact generally reduces cost in comparison to non-technologically driven methods. It does so by creating scale, increasing the amount of work a person can do. In advertising, especially in RFP based buys, we have not seen any significant technologies to drive the productivity of people. Thus transaction costs have not been materially reduced over the course of time. And lastly we have institutional transaction costs. Rent seeking is probably the most well known example. This is where relationships and incumbency come into play. Switching providers, establishing trust, and transacting at scale are all critical functions that lead to policies of procurement and deal making that significantly increase transaction costs.
At the end of the day it all boils down to a simple fact: the lack of standardization in describing advertising assets and transaction rules are the primary driver of RFP transaction costs. The economics will change when standardization takes place. The good guys will benefit and the shysters suffer.