While technology has made re-targeting ads easier than ever (as the Paul Knegten describes it, the practice has become low-hanging fruit), advertisers are now looking beyond this practice for new ways to get new data and to get new people into the funnel. This may be due to duplications in data sets that are ineffective and expensive, or it may be about owning and controlling your own data.
Panel – Beyond Behavior: Mining New Veins of Semantic, Social, and Contextual Data
Paul Knegten, Head of Marketing at Dapper (moderator)
Christopher Hansen, VP of Performance Marketing at 360i
Joanna O’Connell, Senior Analyst at Forrester
Tom Phillips, President & CEO of Media6Degrees
Amiad Solomon, President of Peer39
Chris Stevens, Senior Director of Marketing at Orbitz.com
Stevens believes that practices exist now to bring new customers into the funnel, but it is often cost-prohibitive. Hansen see’s retargeting as a cost-effective catch-all, while more novel techniques may be a crapshoot; they work for some campaigns but not others. O’Connell blames measurement for this difficulty. Stevens questions the efficiency of using techniques like semantic data collection in some areas of the web.
Solomon notes that scale is a shortcoming of retargeting. Often, not enough people will visit a campaign to support the practice. Another technique that suffers from scale, as well as a problem with “creepiness”, according to Phillips, is the influence model. That is, users are often uncomfortable with having their preferences applied to their social graph. Success and failure, according to O’Connell, of such a model depends on who is advertising and to whom. For example, campaign targeted to 18-year old girls succeeded using an influence “peek-in” method.
Solomon explained that display advertising traditionally has not done well with context. This has changed, now that technologies exist to capture the meaning of an article. A limit to this technique, according to Phillips, is that large parts of the web (for example social media or photo sites) are not contextual. Solomon recommends combining different kinds of targeting to maximize your ability to serve an appropriate ad. The problem with just buying data, according to O’Connell, is that they may not get the necessary strategy and technology with it.