Seethroo's CEO speaks about how the company uses Web 2.0 behavioral patterns to target audiences.
Behavioral targeting is proving so popular with advertisers, publishers, and even consumers that it's become a valuable buzz-word in its own right. However, smart advertisers know the "BT" label doesn't guarantee success on the ROI spreadsheet. The central tenet of BT is to give consumers the ad information they really want, so it's not intrusive, not off target, and not too late to influence their purchase decision. Rumors of a new wrinkle in BT have surfaced in connection with a new behavioral targeting and audience measurement company, Seethroo, and have caught the attention of many people in the industry. We sat down with Seethroo's CEO, Jason Frankovitz, to find out what's really new in the world of BT.
Robert Moskowitz: Many people are focusing on different aspects of behavioral targeting. Are you really offering something new under the sun?
Jason Frankovitz is CEO and founder of Seethroo, a behavioral targeting company specializing in user-generated content.
Jason Frankovitz: We think so. Seethroo's approach to BT is focused exclusively on the Web 2.0 arena. Specifically, Seethroo examines the content that individual users contribute to social websites and analyzes it to determine a profile for each person. In other words, we use Web 2.0 activity to learn what each person likes, how much they like it, and when that interest is strongest. Then we use that data to serve an ad that's the absolute best fit for that person's behavior.
Moskowitz: How is this approach different from what else is out there?
Frankovitz: I like to think of our targeting approach as somewhere between "evolutionary" and "revolutionary," not a complete paradigm shift, but more than a minor improvement. There are plenty of companies using methods like content affinity and contextual, and while many of them work pretty well, I think both consumers and clients would say there's room for improvement. We've zeroed in on user-generated content as our profiling vehicle; we feel it's got a lot of potential that hasn't been addressed by what's out there currently, and we can monetize it in a way that's ahead of the curve right now.
Another one of our characteristics is that Seethroo does not carry its own ads; we are not an ad network. Again, there are lots of companies with an abundance of client relationships, infrastructure, and serving technology, and we've no reason to recreate all that. Instead, Seethroo's BT integrates with other ad networks. We help ad networks take the ads they already have and target them better using consumers' Web 2.0 behavior. We don't compete with ad networks, we complement them. So if you're BigAdNetwork.com, and you want to offer your customers a new, premium targeting solution, you talk to us. We're giving ad networks another arrow in their quiver, so to speak.
Moskowitz: How are you defining Web 2.0?
Frankovitz: Our definition of Web 2.0 isn't particularly exotic. For us, it boils down to the difference between sites that control what content to display, and sites that let the audience create and share their own content. Until a few years ago, most websites were essentially loudspeakers, and the site owners were holding the microphone. Nowadays, Web 2.0 site owners are building large, feature-rich concert stadiums, and everyone who shows up can hop on stage and do their thing.
Moskowitz: What kind of Web 2.0 behavior, specifically, are you tracking?
Frankovitz: Seethroo is specifically tracking bookmark activity. We've built a bookmark-based behavioral analysis system (our patent is pending) that examines people's bookmarks, publicly stored on social bookmark sites, to accurately identify topics that are directly relevant to each person. Based on those topics, Seethroo targets an ad that's a good fit for that person. This all happens in real-time.
We also combine all activity across our network to create a "clearinghouse" of bookmark data, which lets us find and quantify behavioral trends that may not be visible when you're looking only at individual bookmark sites.
There are a couple of important things to note about our approach. First, we do not gather usernames, email addresses or any PII from our bookmark sites. That information always remains with the bookmark site and we never see it; this preserves the terms of our partners' privacy policies. We even use random numbers to represent individual people in our system, which further anonymizes all activity. Second, we are an online-only system, we do not analyze bookmarks that people store on their personal computers; we install no "spyware" on anyone's machine.