DSPs’ Next Feat


houdini_small ADOTAS – Demand-side platforms (DSPs) have evolved tremendously since their earliest incarnations in the 1990s. Particularly in the past 18 months, the range of DSP capabilities has exploded. But these are still the early days in the evolution of this space. And, as someone who has been in the industry for over a decade, it’s exciting to know that while we’ve come a long way, we’ve only seen the tip of the iceberg.

To predict how DSPs will evolve, it’s helpful to step back from the acronym soup that we swim in daily (DSP, RTB, MRTB, etc.) and reflect on why they were created in the first place. What fundamental problems are marketers trying to solve such that DSPs, as we know them today, have come to exist?

At a basic level, marketers use DSPs to solve four fundamental problems:

  1. Efficiently reach a highly fragmented audience online
  2. Locate and target the most relevant audience segments, and reduce wasted spend on irrelevant audiences
  3. Deliver the right message to target audiences, at the right time
  4. Continually improve performance on problems 1 to 3

Looking back at history, I think of the buy-side ad servers like DFA (Dart for Advertisers) and Atlas as the first DSPs. They came into existence largely to solve problem 1. Buy-side ad servers brought massive efficiency to what had been an almost entirely manual process. These platforms freed agencies from the challenge of individually creating and managing buys across multiple publishers so that campaigns could be managed centrally by the agency team itself.

Additionally, buy-side ad servers provided solid analytics and reporting capabilities, and thus enabled some relatively efficient learning and optimization. Because ROI tracking and continuous improvement became possible (although not highly automated), we saw a broader base of companies investing more of their marketing dollars in display ads.

The rise of ad exchanges like Right Media’s RMX and DoubleClick’s AdX paved the way for the current class of DSPs. The major breakthrough was that these exchanges allowed unprecedented access to micro-targeted audiences through their APIs. DSPs provided the automation layer, connecting agency buyers to exchange-traded audiences through these APIs, making buying through a DSP dramatically more efficient than purchasing inventory through unique IOs for each publisher buy.

Current-day DSPs have been able to leverage their position to provide capabilities beyond those of traditional buy-side ad servers. Bid management tools provide agencies with much more sophisticated control over pricing. New, more granular levels of reporting help agencies make improvements to their clients’ ads based on a wider variety of performance data. This data also helps them pull the plug faster on poor performing ads as well as justify additional investment in the best performing ones.

Over the past two years in particular, DSPs have evolved significantly, making great progress in solving problems 1 and 2 above. These DSPs provide more and better targeting data, making it even easier to find target audiences online (problem 1) while minimizing spend on irrelevant audiences (problem 2). Agencies and advertisers also increase campaign effectiveness by leveraging automated algorithms to make more informed campaign decisions from the start.

Despite these great advances, a large gap still exists today. Problem 3 — getting the right message in front of target audiences at the right time — is not being addressed by most of today’s DSPs. And that’s not so surprising: creative and media have traditionally been managed separately, and the DSP promise of aggregating audiences across inventory sources and purchasing them efficiently is a pitch directed primarily at buyers.

Until recently, the bulk of the technical innovation has been on the media side of the equation, but many are beginning to recognize the advantages and opportunities of pulling the creative dimension into the mix. Ignoring the creative element effectively leaves billions of dollars of value on the proverbial table.

Take, for example, a conversation I had recently: a marketer described how excited she was to now be able to buy targeted audiences through her DSP technology provider. She and her team were purchasing dozens of different audience segments defined by a combination of behavioral, demographic and geographic characteristics.

On the surface, this all sounded great. But after I asked about the message component, she revealed they only had a handful of different creative treatments to choose from. In other words, each unique audience segment was not receiving a message tailored just for them.

While only a few advanced DSPs provide solutions that encompass both the audience and creative dimensions, I expect many more will follow. When DSPs combine these capabilities into a single platform and process, marketers can cost-effectively micro-target their audiences and deliver distinct ads for each target audience. Problems 1-4 are addressed in a more effective way, and this in turn increases the overall performance of their display advertising.

This is the next major frontier for DSPs.

DSPs that integrate creative and media are the exception today, but even the ones that “get it” will need to continually improve their solutions to stay relevant. In my view, the other major categories of future evolution will be:

  1. Further improvement of algorithms to inform optimal budget allocations and bid prices based on marketing objectives
  2. More and better targeting data
  3. Support for additional ad formats (video, mobile, rich media, etc.)
  4. Ability to bid on/buy “futures” or guaranteed inventory in addition to current spot/remnant inventory

With all the opportunity and interest surrounding display, it’s no surprise that many experts have projected this market to reach $50 billion by 2015. Continuing to solve the four fundamental problems above for an increasingly broad set of advertisers is what will unlock this market’s potential. Consider it the calling of true demand-side platforms.


Please enter your comment!
Please enter your name here