Archive for August, 2010

The increase in buzz and actual growth of the Demand Side Platform (DSP) / Real Time Bidding (RTB) market is not news anymore. The trend of separating audience profiles from media and empowering media buyers  to bid on specific audience profiles across large exchanges of media inventory is a hot topic of conversation, and rightfully so. But of course with any growing trend, it is essential to take the time to identify which players provide true and unique value propositions to the marketplace. Beware of impostors trying to capitalize on the hype rather than helping to perfect the concept of what a DSP facilitates.

In Theory, Practice & Theory are The Same – In Practice They Are Not

In theory, each agency or media buyer needs only one DSP to bid into the entire exchange and second channel media ecosystem, with all the data plugins available at their disposal. The market would have a high degree of bid density (a lot of actual demand side activity) and liquidity (stable supply of replicable “inventory” that establishes and holds its value – which of course is an entire issue in and of itself). Of course, we don’t live in a perfect world, yet. Neither bid density nor an ability to value inventory properly exists in the RTB marketplace.

Give Me a D…

Adding the acronym DSP to your product offering gets more feet in more doors today, and therefore we will see many large networks and new players adding “DSP” to their offerings. However, in concept, many of these new platforms are limited to specific network inventory (albeit large amounts of it), static data profiles or targeting options (albeit fairly sophisticated options), and sometimes lack the total transparency that the more savvy buyers have come to expect from a true DSP (albeit some are willing to work on a CPA basis, so sometimes the buyer doesn’t care). A “true DSP” is one that can bid in real-time into the entire exchange and second channel ecosystem, works with all or at least most of the data providers and maintains total transparency on media and pricing. The holding company-level media agencies have all either developed their own or white labeled AppNexus or other third part technologies. Much like ad servers, as the market evolves, your agency or in-house buyers will only be working with one DSP (or maybe we will start calling them real time bidding engines at some point?) – or at least a primary DSP. Speaking of ad servers – I predict that ultimately Google (DART) and Microsoft (Atlas) will be the two leading DSP’s on the market (although MediaMind will be a third major player, particularly with the impending IPO). This will happen through acquisition, and the first in the category was Invite Media – check one off for Google. Some of the other current acquisition contenders include DataXu, X+1, Media Math, and AppNexus, with new players claiming market entry seemingly monthly. Degree of sophistication of advanced optimization engines should soon become a unique point of differentiation between companies.

Wanted: A Stable Market

Imagine a series of interconnected Venn diagrams, where the overlapping areas represent consumers that satisfy multiple advertisers’ criteria. These criteria are compiled using a combination of data points from several data providers, all integrated into your DSP and available to select from an intuitive  interface. Every single available impression in the exchange is assessed in real-time by every DSP on the market, and multiple bids from the appropriate advertisers within each DSP are all processed in real-time. Those buyers with the highest bids will get the inventory. All of this bidding happens in real-time – billions of times per day. Sound familiar? The bidding part at least? Google built the biggest cash cow in our industry on a similar model – using far less data and sans cross category competition for the same consumer.

Until there is a higher level of bid density and inventory availability, the marketplace will not be ripe for all advertisers and will favor select categories, and not all publishers will provide inventory into the RTB marketplace. It’s the classic chicken and egg problem. Hence some of the non-DSP DSP’s.

The Opposite of DSP

Publishers on the other-hand utilize yield optimizers to interface with the DSP’s to manage inventory, data relationships,  real time bidding and maximize revenue generated. Companies such as Rubicon Project, Ad Meld and PubMatic will soon become necessities for any publisher who wishes to participate in the second channel (inventory not sold at a premium directly by its sales force – which BTW is a lot of inventory!) Even some of these companies are releasing so called Demand Side products. Can they sit in the middle ground of both the supply and demand side worlds? Only time will tell, but my gut says no way. While technology might be able to play both side sof the fence more objectively than people, buyers still want separation from sellers. Of course Doubleclick did it – there’s the whole DFP / DFA thing, but the instances of one company becoming a leader on the buy and sell side of the same technology coin are few and far between.

Anyway, as you can see it’s all really simple …

But in all seriousness, it is as exciting as it is complex. We are participating in the evolution of the digital media world as we know it.

If you enjoyed this post, please consider adding a comment or subscribing to the RSS feed. Thanks.