With many predicting the death of the ad network, 2011 was supposed to be a watershed moment in the world of display advertising. Much of what was trumpeted was that greater transparency could be had by using a Demand Side Platform (DSP)–no more black box–driving greater insights into your audience for smarter buying strategies. Well, 2011 is more than half over, and as far as I can tell, the ad network is still alive so I’m not sure if the promise of transparency has fully arrived.
Transparency can be defined in many ways. Greater transparency would at least mean full disclosure about inventory sourcing, where an advertiser’s message ran, a basic understanding of the algorithms used to identify targets, and identification of the behavioral data that enhanced performance. This information is largely available now thanks to Demand Side Platforms so the question remains: If I have so much transparency, why am I not smarter?
First, let’s remind ourselves of how display advertising has evolved and why ad networks emerged in the first place. In the beginning, display advertising was bought and sold no differently from traditional media — fixed placement for a specific duration of time and for a guaranteed number of impressions. This model is still very much in use and is referred to as contextual advertising. The principal idea is that certain segments of audiences are more likely to be reached on sites that have a higher composition of a desired audience than the overall population. For example, if I’m trying to reach men aged 35-54 with a household income of $100K, I’m more likely to reach my target audience on, say, a business news site than a grocery-coupon site. While this type of buying strategy does in fact help me reach of a lot of the desired audience, inherently there is also a lot of waste; not everyone on a business news site is male, aged 35-54 with a household income of $100K.
As the web began to grow and expand into more niche sites, and advertisers demanded greater efficiencies, the idea of reaching the right audience independent of site or context began to emerge. Performance hungry advertisers needed to deliver audiences in ways that would reach a specific performance objective, typically conversion, in the most cost-efficient manner. Thus, the original ad networks arose to address this challenge. Ad networks’ first clients only cared about conversion, little attention was paid to the factors driving it. Cost-per-acquisition became the ultimate measure of success, and transparency remained in the background.
While ad networks quickly became entrenched with direct-response advertisers early on, agencies with more traditional brand-oriented clients were dealing with a different battle, keeping up with the number of sites and numerous buying options available. Digital planning tools, which evolved from traditional media needs, simply failed to truly serve the needs of the digital media planner. Thus, ad networks began to catch on with these agencies as they were attracted to the planning solution they provided — reaching the audience efficiently–and networks became standard fare in digital media plans. Unfortunately, the success metrics associated with buying via networks didn’t adjust to the needs of non-direct response clients and the “black box” of ad networks continued on.
However, in 2005, the first ad exchange, Right Media, emerged and it appeared that the black box might finally come to an end. Ad exchanges proposed that the more timely the demand could correspond to supply, the greater the value derived, bringing prices in alignment with their true market value. The exchange concept soon incorporated not only media inventory but behavioral data as well, and the promise to eradicate the black box seemed within reach. As we entered 2009, not only had Google’s ad exchange brought significantly more inventory to market, causing advertisers to pay attention, but all the major agency holding companies had formed their own business units built around the idea of forming specialized buying units. Once holding companies came on the scene, the era of the DSP had fully arrived, and again, the promise of transparency seemed inevitable and closer than ever.
It’s mid-2011, and one has to ask if the promise of transparency has been fulfilled. There is definitely more transparency via a DSP. Not only do we have insight into which sites are most effectively achieving our campaign goals, but I can now drill down to which pages on those sites contributed most. Many companies such as Accuen are doing a great job of creating robust user profiles to understand and cluster behaviors to help create deeper customer segmentation. Data providers are also getting better at providing transparency into the recency or frequency of intent data, data “freshness,” and other cookie-level audience attributes. So, with all this transparency, why am I not getting smarter or making better decisions? Has technology perhaps gotten ahead of application? Or, are we even asking the right questions?
Insights derived from the DSPs simply aren’t as actionable as they could be because there is still a fundamental disconnect in the algorithms used by a DSP to find an audience versus the ability to use audience insights to enhance performance. DSP’s were built to deliver efficiency, and until there are better measures of upper-funnel performance that the can be optimized, DSPs will remain stuck in the lower funnel doing what they do best, delivering performance very efficiently. Until there are agreed-upon metrics of success beyond conversion, audience insights derived from DSPs should be used the old-fashioned way by people, and not machines. Additionally, because DSPs focus solely on display advertising and cannot track other types of conversions, the technology only gives marketers a limited snapshot to make critical buying decisions. If the DSP is to evolve and grab more market share, it must begin to address attribution. Until DSPs start pulling in all touch points into the conversion funnel to provide complete transparency into how all media channels interact, the black box will continue.
[Image: Flickr user shannonkingren]