TARGETING
Branding By Behavioral Targeting
April 25, 2005

Study shows behavioral targeting can outperform run-of-site advertising using branding metrics.

Recently, 24/7 Real Media reported on findings from its first network behavioral targeting study. Among the results delivered was increased ROI for advertisers. A large web hosting company, for example, saw a CPA for behavioral targeting impressions 70 percent lower than for sponsorship impressions -- a result of lower CPM, higher click-through and an unchanged ratio of conversions to clicks.

Last week, Revenue Science, Inc. released results from behavioral targeting testing that show increased brand metrics as well.

Dynamic Logic, an independent third-party marketing effectiveness research company, conducted the study on behalf of the Financial Times and NTT DoCoMo. The test measured the impact on awareness and brand attributes of the mobile communications company's campaign running on FT.com.

The primary goal of NTT DoCoMo's advertising campaign, which ran on FT.com from early September through to the end of November 2004, was to raise awareness of the brand among its key target audience, focused mainly on opinion leaders and business decision makers having interests in information technology and telecommunications. The ads were delivered to both an unqualified audience via run-of-site placements and to a highly targeted audience via Revenue Science behavioral targeting.

Dynamic Logic conducted online interviews with those who viewed the ads and the results to a control group that did not see the ads. Both groups were randomly sampled simultaneously while the campaign was live and in-market. While those who viewed the ads expressed greater awareness than the control group, awareness of the behaviorally targeted audience was significantly better; an approximate 200 percent increase in Online Ad Awareness compared to the run of site audience. Results were most prominent in all key metrics such as Aided Brand Awareness, Online Ad Awareness and Aided Ad Recall.

"That last point is significant," says Omar Tawakol, SVP product marketing at Revenue Science. "Often when you dig in, you find a lift in only one data point. But we found lifts on all brand attributes. That really shows how behavioral targeting works."

Awareness Metrics
(Percentage of Awareness by Audience Type)

   Run of Site  Behavioral Targeting  Performance of Behavioral Targeting over Run of Site
 Aided Brand Awareness  56.1%  79%  +41%
 Online Ad Awareness  8.5%  24.9%  +193%
 Aided Ad Recall  10.2%  28.4%  +178%
Note: Differences in performance between BT and ROS groups may also be due to differences in audience composition.
Source: Dynamic Logic AdIndex Oct-Dec 2004, n=723


"FT.com is a premium vehicle for companies like NTT DoCoMo who want to reach our audience of international C-level decision makers," says Shauna Monkman, VP of global online advertising sales for FT.com. "This technology allows us to target groups of people based on what they read rather than by the sections of the site they visit. We are thus able to deliver larger numbers of our premium audience to clients and to customize targeted segments for specific advertising campaigns."

A second campaign objective was to raise the perception that NTT DoCoMo is leading mobile telecommunications development as an operator. Again, the behaviorally targeted approach proved most effective in this case. It increased levels of agreement with all brand attribute statements tested, with results from 45 percent to 83 percent better than run-of-site advertising.

Brand Attributes
(Percentage of Agreement with Statements by Audience Type)

   Run of Site  Behavioral Targeting  Performance of Behavioral Targeting over Run of Site
 Am familiar with company  22.4%  32.5%  +45% 
 Company with potential for growth  27%  45.1%  +67%
 A leader in technology  29%  46%  +59%
 Pioneer in field of mobile communications  26.9%  49.3%  +83%
 Reliable company  21.6%  32.7%  +51%
 Innovative mobile phone operator  29.5%  45.8%  +55%
Note: Differences in performance between BT and ROS groups may also be due to differences in audience composition.
Source: Dynamic Logic AdIndex Oct-Dec 2004, n=723

Of importance to note, says Tawakol, is the fact that these results occurred from a site that already is highly targeted. "Most people look at a site like FT.com and think, 'Why use behavioral targeting? It already draws a targeted audience.' These results answer that question," says Tawakol.

"FT.com came to us with the idea of using behavioral targeting to reach opinion leaders and business decision makers that will be instrumental in growing our brand," says Masanori Goto, manager of international public relations, NTT DoCoMo. "We found that behavioral targeting can be an effective and high-performance tool in conveying our message to the right target groups. We are very much satisfied with the outcome of this campaign."

According to Tawakol, the results are reflective of what Revenue Science has been seeing across the board for its clients' branding campaigns. He notes that they've also seen improved direct response metrics, even for campaigns with branding goals.

Concludes Bill Gossman, president and CEO of Revenue Science: "These results are an important validation of the value that behavioral targeting brings to major, global brands who want to move offline dollars online."

WHITE PAPER LIBRARY

View More Research »