• by Phil Leggiere , September 8, 2006

Behavioral targeting, like all forms of enhanced targeting, is motivated by the holy grail of greater relevance. To that end, all manner of technologies and methodologies designed to extract and aggregate data ABOUT consumers have been advanced. What will characterize the next generation of behavioral targeting, Matt Fleckenstein, vice president of personalization technology firm mSpoke, explains below, is the ability to generate data and feedback from consumers themselves.

Behavioral Insider: What are the most critical limitations of behavioral targeting as it's been practiced?

Fleckenstein: I was really struck at the recent Ad:Tech by a speaker who observed that what advertisers used to strive to do is communicate a value proposition. Now, however, that is no longer good enough. Next-generation advertisers actually needs to gave a value proposition in itself for consumers to even bother with them.... And that suggests a real paradigm shift as far as the whole approach to what targeting is or should be. The shift is from finding ways of pushing what you hope is a relevant message out to prospects you believe should be interested, to making it possible for them to target you on the basis of what they are really interested in. No matter how expertly targeted advertising is, it's not going to realize its potential unless consumers can get the message when and how they want it, and on their own terms.

Essentially, clutter exists when advertising adds rather than subtracts from information overlaod. So the nuisance of poor targeting is that it forces consumers to do the heavy lifting of sifting and filtering out the useful content from the useless. At best, so far behavioral targeting has helped a little by better aiming messages. But what's really going to differentiate publishing and advertising of value going forward is that it gives consumers the tools to better narrow and maintain their own information filters.

BI: Can you outline what you mean by adaptive personalization?

Fleckenstein: Adaptive personalization, as we call it, lets users manage their own content and ad preferences. It begins with the premise that consumers want relevance and want to participate or be proactive in selecting their own advertising, if given the chance. The irony of interactive advertising so far has been that its interactivity has been in one direction only. That's not because of limits in the technology, but because old models of the mass media era die hard.

BI: What sorts of technology underlie the new model?

Fleckenstein: Technologically, two means of personalizing advertising have been developed. One is based on machine learning, which infers consumer interests based on behavioral history; the other is based on consumer input such as recommendations. The first is useful but too black box, the other not scaleable enough. What we've tried to do is create an engine that marries advanced machine learning technology with consumers' implicit and explicit feedback to deliver the most relevant content possible.

BI: How does mSpoke's platform work?

Fleckenstein: When consumers go to the Web site of a content publisher that's using the mSpoke platform, they are asked if they want to create an mSpoke profile. If so, they enter their birthday, ZIP code and a few keywords that outline their primary interests. Based on that information, the Web page and its sections get reshaped to fit those expressed preferences. At that point the platform goes out looking for more articles and ads that relate to those interests. Then consumers can alter or refine their preferences directly in their profile. If they choose to be more proactive, they can pass judgment directly on the items the platform has served up by clicking on little thumbs up/ thumbs down icons next to the content. The algorithm will then use adaptive machine learning so next time a consumer logs in, it will favor the type of content most closely fitting the consumer profile and stated preferences.

BI: How will next-generation behavioral targeting alter the behavior of advertisers?

Fleckenstein: One of the most important performance metrics of the next generation of BT will be about the depth of the ads/campaign--not about the stage of the buyer in the buying cycle.  The next-gen ad targeting will hit consumers up front in the "awareness and comprehension" stages of brand identity/building as well as in the latter "call to action" stages of the buying process. 

BI: Can you cite any examples of campaigns that incorporate a more participatory, adaptive learning approach?

Fleckenstein: I can think of a few current campaigns that really seem to have grasped what participatory advertising and targeting will be about. One that really jumps out is Ford's Bold Moves campaign. There you see something that would have been unthinkable a few years ago, a huge marketer using blogs and blog networks to initiate a conversation with its market, communicating its issues directly to the market and inviting people to comment on those issues. In the process of trying to establish this more horizontal dialogue with the market, Ford is not only acknowledging past mistakes in its environmental policies and performance, but actively engaging consumers' critiques as well as praise. That's an early harbinger hopefully of a mode of targeting based on adaptive learning.