Behavior Trumps Context in Snapple Ads
The campaign, conducted on women's Web site iVillage from late February to mid-April, targeted dieting women ages 18-34 with ads for Snapple's new low-carbohydrate meal replacement drink.
Ads were shown to users who visited the site's Diet and Fitness section at least three times in a 45-day period. The ads also were shown in areas of the site outside of the Diet and Fitness channel.
A Dynamic Logic study found that the batch of run-of-site behavioral-targeted ads outperformed content-targeted placements in branding measures such as ad awareness, message association, brand favorability and purchase intent.
"One of the reasons it works well is that when you're in content, you're dialed into the content so much it might compete sometimes with everything vying for your attention," said Peter Naylor, senior vice president of sales at iVillage.
Deutsch, Snapple's ad agency, ran the ads with the goal of building awareness for Snapple-A-Day, a new product from the White Plains, NY, beverage company. The behavioral-targeted ads represented 15 percent of the iVillage buy.
"IVillage is one of the most premier women's sites," said Patrick Benson, vice president of interactive media at Deutsch. "Inventory is very tight, especially in the Diet and Fitness areas."
IVillage contracts with Tacoda Systems, New York, to offer advertisers audience segments based on user behavior on the site. IVillage offers six preset audience buckets in popular areas: Dr. Mom, Mom as Caregiver, Diet Gurus, Money Managers, New Moms and Work and Money. The Snapple campaign targeted the Diet Gurus segment of 250,000 users.
Dynamic Logic polled those exposed to the ads either in the Diet and Fitness section or on audience-based placements in other areas of iVillage.
Compared with the ads placed next to related content, ads run elsewhere on the site based on users' prior behavior increased brand awareness another 16 percent for all those exposed to the ads; ad awareness, 56 percent; message association, 1 percent; brand favorability, 73 percent; and purchase intent, 29 percent.
"The hard thing is when people have great successes like this, they don't want to talk about them," said Dave Morgan, CEO of Tacoda. "It was important for them to talk about it so people in the industry can know how well behavior targeting can work."
Tacoda, which recently secured $7 million in venture financing, and rival Revenue Science, Bellevue, WA, have deals with dozens of Web publishers to let them serve ads based on site visitor behavior. Such targeting lets publishers charge a premium for ads that appear in lower-priced sections of their sites, while advertisers get better precision at a lower price than content placements.
"With a limited budget, we wanted to make these ads work as well as possible," Benson said.
Snapple's behavioral-targeted ads ran at $10-$15 per 1,000 impressions, a discount from the $25 commanded in the Diet and Fitness channel but a premium from the $5 CPM for run-of-site ads.
"My concern is people are pricing it too low," Naylor said. "I think that's really underselling the market and its value."
IVillage offers behavioral targeting to select advertisers, having run only seven or eight campaigns so far, Naylor said.
"Adding behavioral targeting into an already complex inventory process is very complicated," he said.
Content-targeted ads still performed well. The targeted 18-34 age group's message association was higher in content than behavior, as well as with women who previously consumed a meal-replacement beverage. Still, audience-targeted ads outperformed context among dieters and in online ad awareness among the target group.
"Everyone has focused on context and not on audience," Morgan said. "Just as many times context is the best, there are many times when audience targeting is the best."