When Wall Street's Robots March on Madison Avenue
Since at least 2007, before Doubleclick (bought by Google), Right Media (bought by Yahoo) and AdECN (bought by Microsoft), and arguably for several years before, the advertising industry has witnessed the rise of exchanges and the techniques they have engendered, such as programmatic media buying. Much of this approach was modeled off of practices embraced years earlier by traders of securities on Wall Street, and many of the individuals responsible for bringing them to advertising were veterans of financial services industries.
As these processes have grown in prominence they have served to reinforce a change in preferences exhibited by the literal sponsors of much of the Web's underlying content: large advertisers. These entities historically accounted for the bulk of advertising buys from large publishers (sellers of display advertising), such as AOL, Microsoft or Yahoo, under the premise that qualities of a publisher's brand - driven by the quality of the content on a given site - rubbed off on the advertiser in some way. Under this model, publishers had an incentive to invest in quality content and a consumer experience that could retain consumers on their sites, as the intersection of content and consumers was core to the value proposition a publisher offered an advertiser. But over the past six years, advertisers increasingly focused their buying choices on audiences wherever they appeared, implicitly expressing indifference with respect to the context around which their brand advertising appears. Programmatic buying allowed advertisers to find those consumers more efficiently than would otherwise have been the case, especially as it enabled real-time bidding against individual impressions that are optimized against specific goals.
Programmatic buying has by now taken off - accounting for 20% or more of the display eco-system, and growing rapidly - publishers embracing programmatic selling are forced to make the most of a bad situation. Not selling programmatically leaves publishers worse off, as buyers shift their spending towards inventory that is accessible programmatically. Consequently, many large publishers have embraced this trend, first using programmatic channels to sell inventory that might be deemed as "remnant", or which would otherwise have gone unsold by the traditional sales force. But increasingly, publishers have realized that it may be necessary to make all advertising inventory available in a programmatic manner, through what is now called "programmatic premium". The notion is that a publisher can secure minimum buying commitments from trading desks owned by the agency holding companies or other sources of demand, with minimum pricing agreed upon for certain kinds of inventory. This allows advertisers and their agency trading desks to actually buy inventory using programmatic techniques which they would not otherwise be able to given the mechanics of media buying, facilitating the buying of individual impressions, rather than packages as would be conventionally be bought and sold when humans and IOs (or "insertion orders") are involved.
So costs get cut out of the process for both buyers and sellers, and sellers prospectively gain share in the online display advertising market while maintaining price integrity and limiting sales channel conflicts. What's not to like for a seller? Quite a lot, actually. Our view is that this year's $10mm commitment between a trading desk and a publisher will become $9mm next year, and $8mm the year after that, at least on a like-for-like basis. Advertisers and their agencies are incented to find efficiencies, not to raise spending commitments (except in markets where agencies are compensated by publishers as may more commonly occur outside the United States, but that's a different article altogether). This means that agency trading desks will have a larger pool of potential inventory to "ping" and trade off against as their algorithms look for the most efficient way to drive budgets. At all stages of the buy, the agency trading desk (or its algorithms, at least) has more of a credible ability to walk away from a negotiation with a specific publisher than might be the case in a traditional negotiation. With programmatic buying, whether of the premium or non-premium type, the buyer knows with great speed what inventory is working for them, and can play off sellers who know less and less about the specific characteristics of demand that a given advertiser or trading desk has for a campaign it is executing against.
Given this environment, some publishers, such as Federated Media, have decided that they will be unable to compete by selling digital advertising in the old-fashioned way, and consequently have "bar-bell" strategies of only providing inventory to exchanges, using humans only for the purpose of selling "native" ads which involve a high degree of customization and integration with content. Other publishers, such as Pitchfork Media, have decided to only sell native ads and escape the display-ad rat race altogether. While these companies may be sacrificing some revenue scale, they are arguably setting the stage for a more durable pace of growth in the future, as they can each cultivate their business towards the customers who value their content the most.
Which brings us back to the Wall Street analogy. Over the past decade, electronic trading has dominated the securities industry, and the commoditization of execution has led to a significant decline in fees paid in commissions to Wall Street brokers. According to Greenwich Associates, $9 billion was spent on domestic equities trading commissions over the year ending March 2013, down from a peak of $14 billion for the year ending March 2009. Most of our readers will recognize that the quality of the content historically associated with trading - such as the equity research published by our competitors at sell-side financial institutions with trading functions, and produced as part of a business model that is associated with those trading activities - has generally degraded in recent years, as trading has increasingly shifted to the channels which are most efficient, compressing the fees that were historically paid and which supported the quality content associated with trades: equity research. Essentially, the bulk of the equities trading market (which has been heavily driven by trades associated with algorithms and short-term performance) has failed to express a willingness to pay a premium for their trades because of better insights - or at least, the buyers who represent a sufficiently substantial share of the market are making that expression.
On Wall Street, the preferences of buyers, expressed through
the rise of machines has been a key factor contributing to a smaller market for quality content. We can think of reasons why it won't play out in exactly the same way on Madison Avenue and for
publishers at large, but there will probably be more similarities than differences between the content production businesses and the revenue streams that support them in both industries. As with Wall
Street research, large premium publishers are all but assured of a smaller market in the future than existed in the past. This means that industry participants can benefit as they look towards other
revenue streams (in the case of publishers, e-commerce will become increasingly appealing); big players may be able to take share if middle-sized players fall away; and smaller niche players will
still have a chance to grow as they find new ways to sell their differentiated products for the customers who value them most highly. At least until the robots learn how to produce better content or
write better research themselves.
This was excerpted from Brian Wieser's Madison & Wall report.