|
||
Ed. Note: We at Marketing were excited to have a lively discussion about the role of agency trading desks at our sold-out March 4 Programmatic Trading event. Each Tuesday in our weekly AD-Vantage newsletter, we want to continue that kind of dialogue and free flow of ideas.
In this issue, we share some of the thinking of Brian Wieser, a leading equity analyst who covers the advertising sector from New York. In this lengthy excerpt from one of his reports, Wieser looks at the evolving agency trading desk role within holding companies, including the overstated risk of having their business disintermediated by marketers who establish their own in-house agency trading operations.
Brian Wieser, CFA, is senior research analyst at Pivotal Research Group in New York
The evolution of agency trading desks is important for the agency business, digital advertising, ad tech and – if one is pre-disposed to believe in the notion that all video-based inventory becomes available for trading programmatically– TV, too (although we aren’t convinced about the likelihood of this occurring any time soon).
Agency trading desks are important to the media industry for many reasons. First, they represent a “human” aspect of modern digital advertising which has driven the rise of audience-driven programmatic media buying. By virtue of their position within the holding companies that provide much of the outsourced marketing services activities that brands undertake, these entities have helped engineer real change, encouraging advertisers to prioritize their digital media focus on the audiences that marketers ultimately want to reach rather than exclusively focusing on legacy models of media buying based primarily on content adjacencies.
This human /social engineering factor is not a minor consideration, as so much of what large brands do are dependent upon subjective preferences which typically need to be informed by real people (performance of media is important of course, but there are often many different ways to measure performance or determine what to focus on; this is highly subjective for large brands with meaningful histories and offline presence).
We note that these solutions are often better characterized as programmatic trading platforms. While such a terminology difference is perhaps semantic, it reflects that the trading function increasingly resides with the flagship agency or negotiating unit rather than with what we think of as the trading desk. In other words, Mediacom, Mindshare, MEC and Maxus, not Xaxis within WPP’s GroupM or Interpublic’s Magna Global, rather than Cadreon.
Significant differences exist and will persist between these platforms, such as the degree to which they engage in arbitrage-driven revenue generation akin to the core business model underpinning web 1.0 ad networks or modern-day managed services such as Rocket Fuel.
As an overall consequence, these entities have helped advertisers to express their rising indifference towards legacy models of buying digital media on the basis of content adjacencies. Instead, they have helped brands prioritize their digital media focus on the audiences that marketers ultimately want to reach. Magna Global estimated last fall that $16.6 billion (all $ are U.S.) of digital media would be transacted programmatically in 2014, up by more than double over 2012 levels, amounting to around a third of all display-related digital media buying globally.
Trading desks at the holding companies account for perhaps only a quarter of this amount, but they represent the most significant efforts helping large brands contribute to this growth as they drive more money into programmatic media buying.
Trading desks have had broad impact on the digital media industry. We see them as major contributors to the commoditization the premium inventory sold by the Yahoos and AOLs of the world and also see them as responsible for much of the growth of the long-tail of web advertising, as much of this ad inventory is accessible programmatically via display advertising on exchanges such as Google’s Adx.
The health of the long-tail has contributed meaningfully to overall industry growth and caused much of the weakness we see at many of the web’s formerly dominant properties.
Agency trading desk’s evolving roles within holding companies
The first agency trading desks emerged from within the holding companies’ media divisions around 2007 (and not really on the public scene until 2009) largely in response to the salivatingly high margins that ad networks were reportedly generating paired with the advent of new technologies that facilitated DIY audience building or optimizing media in real-time or driving costs out of media buying (all initially made possible with the emergence of exchanges and tools for buying and selling of ad inventory through them).
Over time, differences emerged between in terms of focus with probably the most extreme distinctions coming from Interpublic (where Cadreon shifted away from a stand-alone business unit and towards a product providing tools accessible by all client teams which facilitates programmatic buying) and WPP (where Xaxis is increasingly taking shape as an end-to-end programmatic trading platform involving the buying and selling of inventory.
Xaxis will view its sibling media agencies as its customers, who are buying from them vs. other providers of inventory. Those media agencies may themselves operate separate automated buying platforms, which could be more accurately be characterized as agency trading desks).
Trading desks at Omnicom (Accuen) and Publicis (Audience on Demand) have reportedly fallen somewhere in-between those of WPP and Interpublic, although their evolution – separately or together – will probably depend to some degree on the shape that the combined Publicis Omnicom Group takes following the completion of their merger later this year.
Relative scale of trading desks against each other and other industry participants
For a point of scale, Magna Global estimated last fall that $16.6 billion of digital media would be transacted programmatically in 2014, up by more than double over 2012 levels and almost 40% over 2013 levels. In total, programmatic trading amounted to around a third of all display-related digital media buying globally.
While not all programmatic media requires a trading desk for a transaction to occur, we think that trading desks have represented the most significant efforts helping large brands to drive more money into programmatic media buying.
Trading desks at the major holding companies captured more than $1B of volume during 2013 as they were structured then, and are posed to continue growing rapidly in years ahead. Estimated billing volumes during 2013 at each of the trading desks can be dimensionalized as follow:
• WPP’s Xaxis: $425mm prior to inclusion of the 24/7 business (the company has indicated total volume of activity for newly combined entity will be $750mm this year)
• Publicis’ AOD: >$400mm in annual billings
• Omnicom’s Accuren: $125mm per WPP estimates
• Interpublic’s Cadreon/Magna Global: Approximately 5% of digital spending, per IPG; $120mm per WPP estimates
Havas and Dentsu operate their own trading desks as well, and all are likely growing at a pace that is broadly around the growth of the market (Magna Global estimates it will be up by almost 40% this year). For points of reference, activity (which might be gross revenue or managed revenue) at Rubicon was $485mm last year, approximately $610mm at Criteo and $238mm at Rocket Fuel, and we would guess that none operate as profitably as the operations within the agency holding companies.
Agency trading desk tactics and competition: Fuzzy and Fluid
If the purpose of the agency trading desk is to facilitate the provisioning of an audience to an advertiser, the ways trading desks accomplish this goal takes varying forms. Trading desks need to help a client to define audiences (using a data management platform involving client-side, agency-proprietary or third party data) and a way to seek out inventory of audiences programmatically, using one or more demand side platforms (DSPs). Any related technologies or data could be licensed or developed in-house.
A range of tactics might then be used to access inventory, arranged via direct access to publisher inventory, access to inventory via supply side platforms (SSPs) who work with publishers, or indirect access via exchanges and networks. They may also work with (and buy from) independent trading desks or other managed services.
Complicating matters further, most of these entities – DSPs, managed services and ad networks – perform effectively the same tasks in slightly different ways with slightly different economics; other entities whose legacy businesses were as exchanges or SSPs (such as Rubicon, notably) also have invested into the capacity to perform all of these tasks themselves rendering distinctions between buyers, suppliers and competitors as incredibly fuzzy and fluid, even to practitioners within the business.
Preferences around costs and transparency often drive the choices that each marketer makes, with performance only one consideration whose importance varies depending on the marketer’s goals.
Marketer-agency disintermediation risks
Many marketers have experimented with or institutionalized agency trading desks inside of their own organizations rather than through the holding companies that buy the bulk of their media. This has occurred in part because of the commoditization of so much of the technology which underpins programmatic access to digital media inventory, because of concerns about opaque pricing and because many marketers by now have experience with buying paid search directly (which is almost all bought in a programmatic manner presently).
Big brands such as Kellogg’s, Ford, P&G, Citibank and Unilever have all been cited in trade press as buying some or all of their programmatic inventory directly, without necessarily involving their agencies. In our view, much of this is temporary. There will always be some large marketers who prefer to manage media buying themselves for the aforementioned reasons, among others (much as there are still some large television buyers who manage their spending internally, too).
Nonetheless, the vast majority will find that the benefits of integration, access to talent and cost advantages even in face of mark-up for services or media found in agency holding companies will ultimately overwhelm any concerns. Further, keeping up with – let alone anticipating or leading – the pace of change in this sector requires a certain depth of presence that few brands can manage in-house given the degree to which this kind of function falls away from those brands’ core competencies.
Consider, for example, the aforementioned Xaxis-24/7 merger, whereby Xaxis is integrating its demand-side capabilities with supply-side offerings which is possible given its ownership (unique among the agency holding companies) of a leading publisher-side ad server. While the absence of such a product combination does not necessarily hold the other holding companies’ trading desks back, it seems more likely to us that trading desks inside of holding companies or independent stand-alone trading desks will be better positioned to evolve than will those controlled by marketers.
Still, some degree of self-sufficiency in this area will persist: web endemics (such as e-commerce marketers) and small brands who weren’t working with agencies previously may very well maintain a durable presence as stand-alone buyers of programmatic inventory through agency trading desks they maintain in-house. However, these marketers, like all the other industry participants referenced previously, will still require a heightened readiness to adapt, as this capacity remains the most critical characteristic of the ever-evolving agency trading desk.