Age and sex don't matter when it comes to TV ad effectiveness, said CBS Corp. Chief Research Officer David Poltrack, who has teamed with Nielsen to create what he called a historic move to replace demographics with a new model for TV planning and buying, based on viewer behavior and attitudes.
The proposed model is a better predictor of what people buy and what makes them buy than demographics ever were, Mr. Poltrack said in a speech to the Advertising Research Foundation's Re:Think 2011 conference Tuesday in New York.
A growing amount of data that matches audience measurement with purchase information shows that using demographics to target commercials is "essentially invalid," he said, "resulting in a misallocation of television advertising investments."
Although CBS paid for the extensive research and analysis used to create the new model, the data and analytics surrounding it will be available to all Nielsen clients, including CBS rivals, Mr. Poltrack said, "in the spirit of open source."
Various parts of Nielsen pitched in, including recently acquired consulting firm Cambridge Group and Nielsen Catalina Solutions, a joint venture that combines data from Nielsen's TV, set-top box, consumer and online panels with Catalina Marketing's shopper loyalty-card purchase data. CBS is also running a series of experiments with advertisers to fine-tune the system, Mr. Poltrack said.
CBS purchased Nielsen Catalina data for 20 categories in health and beauty, household, pet and food products and has studied 15 of them in depth so far. The data confirmed what other smaller studies have shown in the past, according to Mr. Poltrack: "There is no link, none, between the age of the specified demographic delivery of the campaign and the sales generated by that campaign."
Looking at snack foods, only three among the dozens of TV shows most watched by "heavy snacking households" also appear on the list of top shows among viewers aged 18-49, Mr. Poltrack said during his talk, "yet many of the advertisers in this category buy based on age demographics."
So where does the data show a link? "Delivery of heavy users of the product category does have a significant correlation with post-exposure purchase behavior," he said.
In a separate presentation Monday, executives of Nielsen Catalina said ratings demographics by age and sex had a relatively low 0.12 correlation with actual sales produced by exposure to TV ads, where 1.0 is complete correlation and 0 signals no relationship whatsoever. The presence of heavy category users in a TV show's audience, on the other hand, had a much higher 0.69 correlation with sales lift among those exposed to ads on the show.
Mr. Poltrack acknowledged that age in particular hasn't been favorable for CBS. Indeed, in the week ended March 13, for example, CBS nearly matched Fox for overall average viewership in prime-time, trailing by only 10,000 people at 7.97 to 7.96 million, according to Nielsen data reported by TVBytheNumbers.com. But Fox blew past CBS in ratings among viewers aged 18 to 49, with an average 3.4 rating to CBS' 2.1.
Mr. Poltrack added that reliance on the 18 to 49 demographic is hazardous to all media and marketers, partly because it doesn't strongly correlate with purchases and partly because it's declining fast. That group's share of U.S. population declined from 62% to 57% between 2002 and last year, he said, and will drop another two points to 55% by 2016.
Nielsen Catalina can produce TV ratings for heavy detergent users, Tide enthusiasts or brand switchers as an alternative, but its executives acknowledged it wouldn't necessarily be easy to create programming around that. And then there's the complexity of potential TV deals built around thousands of potential category and consumer-behavior segments.
To simplify things, CBS turned to Cambridge Group to create a six-part segmentation model that uses behavior, motivation, programming preferences and psychographics. Kevin Bowen, principal at Cambridge, said this segmentation model could improve delivery of gross ratings points of heavy category users or other target audiences by 20% or more, compared with going after age-sex demographics with similar ad spending. Better matching ads to consumer interests, he said, also led to substantial improvements in ad, brand and message recall as well as ad likability.
The six segments in the new CBS-Nielsen order are:
TV companions: For this group, TV is almost always on and is like a member of the family.
Media trendsetters: Early adopters of technology and new content, and also 39% multicultural.
Sports enthusiasts: Made up mostly of men, but most guys aren't classified here. This group also likes action-adventure programming.
Program passionates: Highly involved with favorite shows, and the biggest DVR time-shifters.
Surfers and streamers: Most open to watching alternative content on TV and most often using laptops or tablets to multitask while watching TV. They skew young, but include a large component of 50-plus people.
TV moderators: Those who enjoy being experts and leading others' choices.
Mr. Poltrack said the new segmentation model and research culminates work he's been involved with since 1992, and added that CBS is working to apply it to digital, radio, print, out-of-home and local markets as well.
The growing use of single-source data like that from Nielsen Catalina and TRA Global, which combines set-top box and shopper-card data, has started to have an impact even before this, Mr. Poltrack said. That data led marketers to restore advertising budget dollars -- to TV in particular -- faster in the recent recession than in prior ones, he said.
"Marketers quickly reversed course and increased their advertising investment even before the recovery was really under way," he said.