• by Kate Kaye , August 17, 2004

Targeting users based on previous and real-time Web behavior boosts click-through rates and conversion rates, and enables advertisers to reach desired consumers with fewer ad impressions. So says a June 2004 study conducted using Advertising.com's Optigence research platform.

The "Behavioral Targeting: Improving campaign performance through customized behavioral segmentation" study analyzed performance of a total of 500 million ad impressions served by three distinct advertisers employing Advertising.com's LeadBack behavioral segmentation technology. Like other behavioral targeting products provided by companies such as Dotomi, Poindexter Systems, Revenue Science, and Tacoda Systems, Advertising.com's technology targets ads based on anonymous user interaction data. Specifically, LeadBack works by using non-personally identifiable information gleaned about users on an advertiser's Web site to target ads to those users across its network. According to comScore Media Metrix, the Advertising.com network reached 118 million or 76 percent of U.S. Internet users in June, topping all other networks.

"Knowledge of previous interest in a specific advertiser's offerings has an enormous impact on an advertiser's campaign," suggests John Ferber, chief product officer at Advertising.com.

The first advertiser's campaign, which took place over a two-week period in June, was intended to propel users to complete an online registration form. A substantial 360 million ad impressions were served--both to users who had been identified by the system as having previously visited the registration Web site but not completing the form, and to non-targeted users across the network. For all three campaigns studied, ads were delivered across the entire Advertising.com Web network--and, according to Ferber, ad creative was identical for both behaviorally targeted and non-targeted ads.

The study determined that behaviorally targeted impressions from the first advertiser resulted in a 192 percent rise in click-through rate and a 167 percent boost in conversion rate. In addition, the effective cost-per-thousand impression rate for behaviorally targeted impressions rose 167 percent, from an average of $1.49 for non-targeted impressions to $3.98. The higher the effective cost-per-thousand impressions rate, the fewer the number of impressions it takes for an advertiser to reach targeted consumer segments.

The second advertiser ran about 21 million ad impressions from July 28 through August 10 in the hopes of spurring credit card sales. The system isolated users who had visited the client's landing page, but did not make a purchase. Compared to non-targeted impressions, those served using behavioral targeting garnered a 94 percent rise in click-through rates, a 2,232 percent increase in conversion rates, and a 522 percent jump in effective cost-per-thousand impression rate. According to the study, the advertiser essentially earned $3.44 per one thousand behaviorally targeted ads compared to $0.55 per one thousand non-targeted ads. The Advertising.com study is "just another study that reinforces that the behavioral targeting approach works," comments Peter Naylor, SVP Sales at iVillage. The publisher uses Tacoda Systems' behavioral targeting technology to identify audience segments like expectant moms, fitness fanatics, and auto buyers, and to target ads to them no matter which content area of the site they're visiting. The third advertiser included in the study aimed to acquire new customers, and sent about 140 million impressions from July 26 through August 8. Behaviorally targeted ads were served to users who had visited the home page, but hadn't signed up for the advertiser's service. Ads served through the LeadBack system resulted in a 225 percent click-through rate increase and a 3,130 percent conversion rate increase as compared to non-targeted ads. The effective cost-per-thousand impression rate went up 2,978 percent as compared to non-targeted impressions. According to iVillage's Naylor, there are few advertisers--mainly those looking for broad audience reach such as movie advertisers--that are not well-suited for behavioral targeting. "It's good for publishers, and it's good for advertisers," he concludes, adding: "If someone's identified as being in the market, and you know that and put ads against that, it's going to work."