Behavioral Insider: One of the areas you talk about in your 2008 trends predictions is the semantic Web. You say that while this technology has largely lived in the research arena so far, it has finally matured and is now ready to make its way into mainstream applications. How do you see marketers starting to leverage the user data from tools like Twine, Yoono, del.icio.us and Stumbledupon.com?
Alistair Goodman: I'd say of all the top trends for 2008 we identified, the semantic Web is the most 'bleeding edge.' But we think there's a huge opportunity for advertisers to focus technology that's already here to increase their knowledge about how consumers actually interact with content. In our case we use technology on content pages to deduce the true topics of pages on levels far more precise than keywords. We've been focused more and more on how to use page topic knowledge to decide in real time not only which products -- but which ads -- will be the most relevant.
BI: You seem optimistic video may become much more targetablebased on user attributes.
Goodman: Yes. With video, I think the frustration stems from the complexity of the buy, as marketers have gotten much more power about how and where their display ads are deployed. What's lacking now are standards and simplicity in serving video. But one major priority of the coming year will be to extend the full gamut of targeting options they already have for banners.
BI: Having local advertisers become able to target national sites is another Exponential trend. What's your road map for that?
Goodman: We're thinking that 2008 in many ways will be the year hyper-local targeting becomes feasible. The real opportunity in local targeting is to allow a local advertiser to run an ad on a national site and reach just those visitors on the site from specific zip code areas. Within specific geographical areas you could then segment by behavior. Of course if you're only talking about three or four zip codes, scale will be an issue. But I can see regional advertisers making great use of something like that. What this opens up also is the ability to dynamically change content based on both behavioral and geo segmentation.
BI: So far one-to-one personalization of creative content has been a very expensive proposition, mostly affordable by only the larger retailers. How and why do you see this expanding in the coming year?
Goodman: We've begun to see one-to-one customization of creative content and product merchandising beyond the large retail space. We have a large insurance firm, for instance, now changing product and content mix based on past habits and current activity of visitors. Utilizing dynamic content controls your customize creative by demographic and behavioral profiles, not just in terms of offers but by aesthetic components that convey different aspects of brand association and image.
BI: Targeting for influence rather than purely transactionally is another goal that's been long talked about. Why do you think this year the effectiveness measurement will move decisively beyond clicks?
Goodman: Increasingly we'll see a greater emphasis on knowing more about who views campaigns but doesn't necessarily click. There's a world beyond click behavior that we're just beginning to truly understand enough to get some leverage. Part of that is learning to understand the effects of a campaign on the particular demographic, psychographic and geographic targets of the campaign. Campaigns could have impacts on awareness, image, interest and many things that influence subsequent behavior but aren't encompassed in a click.
Even more interestingly, we're starting to look at the impact of campaigns on audiences you didn't target. Identifying surprising, unpredictable behaviors you would have missed in the past is as important as leveraging known behaviors. For instance, what if your campaign is geared to males 18-34 in cities, but it turns out the campaign is actually seeing a lift among 35- to 49-year-old males in the suburbs? If you know that in a timely enough fashion, you can shift your focus and direct new offers and approaches based on those insights.
The final thing I'd say looking ahead is that we're now at a point where we can finally begin leveraging all the data assets we have in real time, based on up-to-the-minute information. In the past collecting, storing and interpreting data was a distinct and isolated phase. You crunched and analyzed the data. Then, in a very separate process, you tried to take whatever insights you could glean and develop a creative to translate them into action and execution. Both processes were on different tracks. For the first time, they can now truly begin to converge.