For all of our adult lives, we’ve been increasingly sensitized to the notion that we’re in a demographic — the males between 18 and 34 demo, the females between 34 and 49 demo, the over 50 demo, what have you. The notion that age and sex correlate with audience behavior has been a hallmark of media advertising for decades.
Where did this come from? From a time-honored battle of networks. Historically, CBS was the dominant television network, pulling down Nielsen ratings none of the other networks could compete with, until an ABC executive had a bright idea, according to Neal Gabler in “The Tyranny of 18 to 49“:
CBS’s ratings hold was impregnable, but only if one measured the whole audience. It was the genius of Leonard Goldenson, the head of ABC, which routinely finished in third place well behind CBS and NBC, to change the rules. Unable to compete in the big sweepstakes, Goldenson created a smaller one that he could win. ABC, he told advertisers, was getting younger viewers than CBS and NBC, which meant that advertisers with products appealing to that segment would be wise to buy time on ABC. That was akin to a football team saying that it had lost the game, but it had gained more yards. . . . The network had managed to make itself seem successful despite its consistent third-place finishes. It was all smoke and mirrors.
And so was born the notion of slices of the audience based on age and sex, with some slices more valuable for certain product categories than others.
Now, in a long-overdue response to ABC’s maneuver, a CBS executive, the appropriately named CBS Corp. Chief Research Officer David Poltrack, is coming back with a game-changing proposition — age and sex don’t correlate with purchasing behaviors. In fact, he called them “essentially invalid, resulting in a misallocation of television advertising investments.”
How poorly do age and sex measure purchasing behavior? If they correlated exactly, we’d have a measure of 1.0. Instead, Poltrack and Nielsen have measured them as having a correlation of 0.12, much closer to the null value of non-correlation.
This is not good news for the paradigm of demographic marketing and advertising sales.
Poltrack is teaming with Nielsen to create a new set of measures based on viewer behavior and attitudes, which they claim show a 0.69 correlation with purchasing behavior.
How does this matter to scholarly publishers? Some of us sell advertising. When the paradigm shifts with the big agencies and buyers, we feel it.
Potential winners may be those who can measure behavior and attitudes. Of course, as one blogger notes, these kinds of measures aren’t new to online:
Apart from a few situations (like targeting 18-19 year olds because you know many of them are preparing to head off to college, for example) age based demographics ARE completely useless. Social media and digital advertising has already been leading the way with methods for buying media that are inherently more focused on behaviour (keyword text advertising) or attitudes (targeting by interests on Facebook).
However, for advertisers accustomed to buying print media, and journals accustomed to selling it, breaking the habit of selling to demographics — and creating the infrastructure to allow for better targeting by attitudes and behavior — will take time and effort.