It became politically incorrect long ago to spout assumptions about what people do or believe based on their gender or age. Yet those demographics still rule a big part of the marketing universe, where hundreds of billions of dollars of media still trade in buckets labeled things like "women ages 18-to-49" on the assumption that this is a critical driver of what people will buy.

True, a growing number of media analytics firms segment audiences hundreds of other ways—finding beer buyers, new parents or people who recently moved. But Nielsen’s age and gender demos still rule TV deals, and often extend to measurement in digital and other media as marketers seek cross-platform consistency. So while marketers and agencies have moved beyond the old demos rhetorically and practically, they also can’t fully shake them.

Chalk it up to a combination of tradition, inertia, and a track record for a measurement system that, despite its limitations, has years of validation. While digital media is largely traded in far more complex audience segments or individual targeting, linear TV and other traditional media can still generally be measured only using data collected from consumer panels too small to handle that much detail, points out Jane Clarke, CEO of the Coalition for Innovative Media Measurement.

“The media industry unfortunately moves like an oil tanker in a way,” says Media Rating Council CEO George Ivie. “It’s difficult to turn. We’ve been working with age and gender for a long time. All the vendors are enabled on age and gender. All the vendors that have captured currency status generally use that.”

Marketers ready for change

Even so, the market is ready to move on.

“Demographic audience profiles are an antiquated system,” says Procter & Gamble Co. Chief Brand Officer Marc Pritchard. “We’re really moving to a very consumer-centric type of media system that really is driven much more by consumer behavior and content.”

That includes relying on a first-person database with 1.5 billion plus consumer IDs and data signals to know when people are ready to buy a brand or a category. So, for example, Olay starting two years ago used a “smart audience” targeting approach that seeks high potential users rather than “women ages 25 -to-54,” and spent 10% less while reaching 20% higher sales growth.

The trouble is, smart audiences are most readily found in digital media, and even then, they can be harder to track reliably.

Age and gender are fixed or at least very slow-moving targets, whereas brand and category product buying can change dramatically depending on people’s circumstances or whims of taste, says NCSolutions Chief Research Officer Leslie Wood. “That gives age-gender a lot more stability when you’re measuring audiences,” she says. “And I think most content is kind of directed to age and gender.”

All that makes it simpler to buy and sell media using Nielsen demos, she says. Yet purchase-based segments are far better at directing advertising to people who will act on it, she says. As a result, she believes NCSolutions purchase-based audience targets are used commonly in media deals, even in TV, where Nielsen gross rating points may be the nominal currency. It's not quite like buying an addressable audience, but at best Wood says, a particular buyer segment, like pet food buyers, may be 20% to 40% more common in one TV program than another.

An executive with a major media agency group, who spoke on background, said the market really has evolved toward a two-tiered approach. Agencies may be doing TV deals with networks using Nielsen age-gender demos, but they’re using more advanced audience metrics to optimize what inventory gets parceled out to which brands.

“The questions are what are you playing against and what are you negotiating against?” the agency executive says. “It’s not all one thing.”

Josh Chasin, chief measurement officer of TV and video analytics firm VideoAmp, agrees. “I think a lot more planning and buying is happening against advanced targets than is realized,” Chasin says. “Agencies are doing it but not telling the media companies they’re doing it. They perceive an advantage in the transaction if they don’t show their cards.”

So, for example, audiences age 50 and up may have relatively little value in a media deal, but far more for many brands. Agencies get the less favorable demos cheaper. Then they apply analytics services to find where their more advanced audiences over-index on linear TV, he says.

Fairness and reliability issues

But buying with age-gender demos in one hand and advanced analytics in the other also creates a couple of other problems.

One is that agencies aren’t really being compensated for the value they provide, the agency executive says. “Brands are making media pitch decisions based on things like price and payment terms,” he says, and often the only way to objectively measure that is how well an agency delivers the lowest price for Nielsen age-gender ratings. The industry needs to evolve to compensate everyone based on the more advanced audiences they’re really delivering, and ultimately how well those media buys produce sales results, he says. “The measure for success shouldn’t be based on exposure through an antiquated survey system and more about business outcomes.”

The other problem is that the firms measuring more advanced audiences or business outcomes aren’t nearly as well vetted as Nielsen’s age-gender demos, because they aren’t audited by the MRC.

“People feel like this is new anyway so maybe we don’t hold them to the level of transparency and rigor that the MRC does. And for some reason they’re surviving in that area,” Ivie says.

“I think the marketplace has been talking about forcing more auditing in this area. There hasn’t been anybody come to the table to start the dominoes falling. But I feel it’s necessary to legitimize more advanced targeting. Once one of them falls, all of them are going to have to fall.”

Beyond segmentation to outcomes?

Potentially, that’s a lot of audits and a lot of payments to the MRC. And maybe one solution is to skip right past vetting hundreds of audience segment measurements in favor of simply measuring how well media deliver the ultimately desired outcome—sales. Trouble is, the whole ROI measurement side of the industry isn’t third-party accredited yet either, and the standards for that haven’t even been written.

Ivie points to the plethora of glowing studies showing positive sales or brand perception lift straining credulity and creating a need for some policing. “All these studies are telling you every time you run an ad campaign you get brand lift,” he says. “But advertisers aren’t growing that much.”

The MRC has begun developing standards for outcomes-based media measurement, which Ivie expects to emerge sometime next year. After the standards comes the work of auditing and accrediting potentially dozens of firms that do ROI and similar measures of business outcome like lifts in brand favorability. But that’s realistically years away.

In the meantime, that leaves age-gender demos as the most vetted measures of audiences.  And Nielsen believes they're going to have a role for a long time.

Having its cake, and icing it

“I look at it as a layered cake,” says Scott Brown, general manager-audience measurement, at Nielsen. “The bottom layer, which I don’t think is going away, is age and gender. Everyone has an age and gender. And I think brands will always want to understand, broadly, who are they reaching. That is always going to be the base foundation.”

But then even Nielsen is adding new layers and some icing, including its recent deals with AT&T’s DirectTV, Dish and Vizio to use their 55 million devices to segment audiences well beyond age and gender. That opens the door for a lot more TV deals based on addressable audiences of cat lovers, beer drinkers, fast-food aficionados and the like.

But that system, doesn’t start producing data until next year and could be years away from MRC accreditation or production of enough date to let networks significantly increase their sales of advanced audience segments.

Decades ago, well before digital media, marketers were buying magazines based on Simmons Research segments that were much more refined than Nielsen age-gender demos, yet still having to hunt for those same segments on TV by finding demos where they over indexed, says Bill Livek, CEO of ComScore, which has its own addressable TV offering that’s already producing data.

“Moving away from age and gender is just moving away from a custom that for decades wasn’t working, and we were all kidding one another,” Livek says.

Ultimately, the death of the demos may come when sellers realize that better measurement probably makes their media more valuable in total, not less, Chasin says. “There’s always been a trend toward more granular measurement, and sellers are always afraid of it. It does upset the status quo, but historically it’s always resulted in a rising tide.”