“Data Driven Thinking” is a new column written by members of the media community and containing fresh ideas on the digital revolution in media.
Today’s column is written by Pascal Bensoussan, VP of Products at Aggregate Knowledge, a buy-side optimization platform.
In 2006, Greg Stuart, Interactive Advertising Bureau CEO, estimated that advertisers’ waste, that is, impressions reaching the wrong audience or none at all, was $112 billion a year in America and $220 billion worldwide, or just over half of their total spending. Wanamaker was remarkably accurate.
In 2009, we have made little progress. A recent research from ComScore indicates that only a small minority of campaigns reach their intended audience with the desired frequency. Of eight U.S. brand campaigns with budgets between $400,000 and $2 million, ComScore said none reached more than 20 percent of their target with a frequency of four impressions or less. Up to 80 percent of the impressions were delivered either to the wrong segments in the U.S., or to consumers outside the U.S.
A number of factors are to blame for this situation:
1. Reaching audiences based on context is expensive and not particularly effective. Yet, media planners and media buyers are only trained to buy audiences based on the media they consume. It takes a fair amount of interpretation and mapping to turn the target audience defined in a creative brief into a robust media plan. Emerging technologies enabling multi-party cookie synching, real-time decisioning, real-time bidders, and media de-risking with new classes of publishers and ad networks, provide effective means to buy the right audiences directly instead of guessing the web properties they might visit.
3. Many media agencies are still in the Stone Age when it comes to digital media execution. They are not to blame as much of the technology innovation of the past five years was focused on the publishers. The lack of automated tools makes it hard for media planners to create many mini-media buys optimized for a more granular and more precise set of audience segments. As a result, media planners will typically focus on larger, less accurate media buys that can be executed quickly and with no hassle from a limited, but known, set of inventory providers.
4. A campaign is often like a house of cards relying on proprietary code to synchronize message passing between all parties involved. The lack of standards in ad operations makes it hard to coordinate data collection and media buying flows across the complex network of publishers, networks, exchanges, and data providers. The proper orchestration between retargeting pixels, ad tags, user pixels, third-party tracking beacons, and click redirects can be daunting. A single error can have dramatic implications on the ability for a campaign to reach its desired audience.
5. Trafficking errors are introduced throughout the value chain, either advertently or inadvertently, especially when ad networks sub out media buys to other ad networks, a common practice. In such cases, even when a problem is fixed somewhere, the fix is not passed along to downstream networks and publishers. Campaigns explicitly targeting users in the U.S. may see 5 to 20% of their impressions outside the U.S.
6. Many data providers only offer a limited view of the user. In a recent campaign targeting home-based business users with children, we had to combine micro-business users, home-based users, and households with children from three separate data providers. Scaling audiences can be particularly difficult without the proper infrastructure to normalize, aggregate, and store audience profiles across providers.
7. User cookies are misleading. A cookie is unique to a computer, not a person. The information stored in a cookie typically aggregates behavioral clues from multiple users of different gender, age, and interests. Some data providers claim to support cookie-level targeting when, in fact, the data stored in the cookie relies on, sometimes dubious, transformations such as IP-to-zip code mapping or mass inferences from small panels. The key to this challenge is to triangulate and leverage data across multiple sources, in near real-time.
8. Hitting the same user 20 times in the same week is like hitting the wrong user. In his recent post on iMedia Connection, Michael Estrin quotes Guy Schueller, media director for Organic saying “The sheer number of networks and subsequent overlap of inventory has created frequency issues and, at times, confusion.” Many campaigns end up recycling the same pool of targeted users across their data and inventory providers. Without real-time visibility on users reached and inventory overlap, agencies are unable to manage frequency capping waste and prioritize spend toward the most efficient providers.
9. On target audience does not mean the right audience. Most publishers and ad networks provide little visibility over data and optimization techniques, making it difficult for media buyers to understand the hidden characteristics of performing audiences that offer new branding opportunities or better direct response performance.
10. Little control to optimize in real time. The media mix of most campaigns is only optimized once or twice through their flight, at best weekly. Advertisers have little control to optimize the audience reached by their campaign. As media buying becomes more transparent, more efficient and real time, it will be possible to modify media buying and audience targeting strategies dynamically, optimizing not just against a set of audience segments, but against frequency, day parts, context, and messaging.
The real-time media buying, audience management and campaign optimization capabilities of emerging demand platforms finally give a voice to the advertiser where previously the whole ecosystem was dominated by ad networks and publishers. Those platforms promise greater accuracy, scale, and efficiency to reach the right audience with enough transparency and control to let the buyer decide when, how, and at what price to do it.
Follow AdExchanger.com (@adexchanger) on Twitter.
