A growing number of TV buys are tailored by using loyalty-card or other databases to target people based on what categories they actually buy rather than demographics. But research from Nielsen Catalina Solutions, Starcom MediaVest Group and Coca-Cola Co. suggests going beyond category buyers to target a narrower group of likely brand buyers can produce far better results.
In research released June 15 at the Advertising Research Foundation's Audience Measurement 2015 conference in New York, NCS and SMG outlined a complex method for targeting those likely brand buyers that could produce significantly greater sales lift at a lower cost than just aiming for category buyers.
Coca-Cola's Simply Orange brand participated in the study, which was also sponsored in part by CBS. Nielsen Catalina brought in a dozen other packaged-goods brands to round out the analysis.
Using simple demographic targeting of TV ads could deliver the 50% of households who accounted for 57% of incremental sales generated by ads for Simply Orange. But focusing on a more refined buyer target could deliver a smaller 31% of TV households that accounted for 72% of incremental sales.
That target included a mix of people who were buyers of the brand before seeing the ad, as well as some who weren't. Mining Catalina shopper data and Nielsen viewership data found people most likely to respond to the ad were in households demonstrating what NCS Chief Research Officer Leslie Wood called "choice fragmentation." These households tend to switch brands more frequently or buy multiple brands, perhaps because they prefer variety or have household members with differing preferences, Ms. Wood said in an interview. They are, in other words, a bit fickle. Targeting such "likely brand buyers" in some cases found a group five to 10 times more likely to buy as a result of seeing an ad -- in the case of baby care or soft drinks -- than other category buyers, she said.
Just as media agencies have to think beyond the traditional media mix, they also need to look at varying the "data mix," said Jeff Chaban, senior VP-business integration and analytics for SMG. That means getting data to help define the target, not just find it. "It's really setting the pace for how we're going to think about audiences going forward differently," he said.
By "re-optimizing" a TV schedule to aim for the ideal Simply Orange target, SMG found it could have increased audience reach by 2.2% at a 17% lower cost vs. simple demographic targeting.
In a statement included in the presentation, Coca-Cola said the study showed "potential for significant reach and cost efficiencies," and that the next step should be further testing with actual buys to prove the targeting approach delivers sales more efficiently.
CBS Chief Research Officer David Poltrack outlined how in the cases of snack, oral care, dessert or other brands, adding just four additional network TV spots to existing schedules of hundreds could significantly increase the proportion of target buyers reached.
Most buys today are at least somewhat informed by overlaying analyses of viewer purchase patterns on the traditional demographics against which deals are written, Mr. Poltrack said. As a practical matter, it's impossible to develop programming specifically for people open to switching orange juice or toothpaste brands. But CBS does have a seven-part segmentation model aimed at delivering audiences such as "program passionates," who are interested in healthier food and high-quality drama, and those tend to help produce high concentrations of various brand buyers too, he said.
"We don't just test our programs with adults 18-to-49," Mr. Poltrack said. They're also tested to do well with "program passionates," "media trendsetters" and "sports enthusiasts," each of them groups who tend to be big buyers in categories important to advertisers.