GOP Races To Catch Obama 2012 Campaign Data Techniques

GOP Races To Catch Obama 2012 Campaign Data Techniques

On August 21, Florida Gov. Rick Scott’s campaign purchased an ad on the CBS reality show Big Brother for roughly $6,000. The ad, tracked by ad monitoring firm Kantar Media, ripped into Scott’s opponent, Republican-turned-Democratic nominee Charlie Crist over alleged corruption and was seen by approximately 19,000 likely voters.

The same day, Crist ran an ad on “The Doctors,” a Dr. Phil spin-off featuring a team of medical doctors discussing health-related issues, narrating a self-congratulatory tale of his career and vowing to “fight for you.” The ad cost $50 to run and was seen by about 3,253 voters.

The Scott ad was seen by about six times as many voters, but cost 120 times more, meaning that for each voter it reached, Scott spent 31 cents, while Crist spent only a penny, according to data provided by a GOP consulting firm called 0ptimus.

The data is on the cutting edge of new methods about analyzing television audiences, with the key innovation being a reliance on viewing data from hundreds of thousands of cable “set-top” boxes rather than the traditional Nielsen ratings that are based on measuring the viewing habits of roughly 20,000 households nationwide.

Pioneered by President Barack Obama’s data-intensive campaigns in 2012, GOP firms like 0ptimus, Deep Root Analytics and i360 are developing sophisticated ways of estimating the most efficient shows to purchase political ads on, and the results can sometimes be surprising.

For example, in a recent campaign, 0ptimus found that “Queen Latifah” daytime talk show was a surprisingly cheap way to reach Republican voters.

In the 1970s, many campaigns relied on traditional advertising firms to place ads. Robin Roberts, the co-founder of industry behemoth National Media, is credited by some for creating a service for campaigns unique to the political sphere.

But like other industries, the campaign advertising sector is in the midst of a giant upheaval from new “big data” techniques that are giving politicians the ability to target specific slices of voters in more accurate ways and even buy ad placement in specific homes.

“I think it’s a tipping point. The campaigns of the 2016 presidential nominees won’t really use Nielsen at all, except indirectly through rate cards when they’re looking through the rates,” said Elizabeth Wilner, a contributing editor at Cook Political Report and senior vice president at Kantar.

One leading 2016 contender is already taking notice: Sen. Marco Rubio (R-FL), whose super PAC has already paid 0ptimus $200,000 in the 2014 election cycle, according to the Center for Responsive Politics.

That’s a significant amount this early in the 2016 campaign, but it’s a drop in the bucket compared to what a presidential campaign spends on buying ads. For instance, during the 2012 presidential race Massachusetts Gov. Mitt Romney’s campaign paid $234 million to American Rambler Productions, the firm it set up to place ads that was criticized for spending inefficiently.

The traditional way of placing ads for a campaign to flood the airwaves is to buy 1000 Nielsen points per week per media market.

“Gross ratings points,” or GRPs, as they are referred to in the industry have been a mainstay for decades. An actual ad placement proposal from a TV buyer obtained by Breitbart News shows how simple the plan is: each market is listed, along with the GRPs proposed by week (maxing at 1000 in the final weeks of the campaign) and how much it would cost to purchase the ads based on the average price.

There’s no discussion of targeting specific voter segments or difference in audience or price for different shows.

0ptimus’s method, which they have detailed in conjunction with releasing data from the Florida gubernatorial race as a case study, is to build one or multiple voter models based on high-sample polling, retrieve a large sample of those voters’ viewing habits as recorded by set-top boxes (the DVR or channel control box) and calculate how many targeted voters are watching each show.

The data comes in daily, in 50gb text files sorted into bins of voter groups and tracking each household, by an anonymous id number, in 15 minute increments.

Comparing the ratings and prices (from “rate cards” from the television stations) with which ads are running – data available from Kantar – gives 0ptimus the ability to analyze how many people each campaign is reaching and at what cost, which they have tracked and charted over the past seven weeks.

The firm is also tracking support for each candidate, showing that Crist has seen a slight uptick in support in the past several weeks despite the fact that Scott is reaching more voters – and at slightly lower cost per impression than Crist.

According to the data, either campaign could be spending significantly less (roughly 3 cents per impression) to reach about the same number and type of voters.

In the 2012 campaign, Obama’s team used the same approach to target a list of 15 million persuadable voters in swing states, ultimately purchasing twice as many lower-cost cable ads as the Romney campaign on twice as many channels. The campaign’s analysis led them to purchase ads on relatively low audience shows like “Judge Joe Brown” and 1:00am repeats of “The Insider”, which could be bought cheaply.

Romney’s campaign didn’t use the same approach, prompting criticism about how hundreds of millions of dollars were spent.

Top adviser Stuart Stevens told the New York Times it didn’t make sense for what Romney had to accomplish. “We still were trying to inform likely voters who Mitt Romney was,” he said, “and until you get those voters, you’re insane to go off and say, ‘We’re going to try to win this with left-handed Lithuanians.'”

There are other differences as well, Brad Todd, the co-founder of On Message, Inc., a leading GOP ad firm that works in close collaboration with National Media and is producing Scott’s ads in Florida, noted that Democrats watch more television than Republicans because they’re home more of the time. Obama’s campaign was often trying to motivate individuals with a low likelihood of voting to show up on election day, perhaps the types of voters who might be more likely to be watching re-runs of The Insider at 1:00am.

Todd also says reporters, who are often ignorant of many of the intricacies of the industry, have hyped the Obama story.

“The Obama campaign is really good at getting stories about how great the Obama campaign was. Their core competency is getting good publicity,” he said, adding that he wasn’t criticizing any specific tool, each of which has a place for different situations.

Evangelists of the new methods say the gains can be staggering.

“It’s absolutely a huge deal. This level of analytics for television has never been available before,” said Chris Wilson, a senior executive at Rentrak, a media data company that is working to invade Neilsen’s traditional turf with the new data.

Scott Tranter, a senior partner at 0ptimus, said “Generally speaking, set-top box data costs between 1-3% of the total ad buy and we’ve seen efficiency gains of between 5-20% depending on the state and the media market.”

To put that in perspective, five percent of the $234 million the Romney campaign paid American Rambler is $11.7 million dollars.

Another technique being developed is advertisements targeted at specific households, called “addressable” ads in industry parlance.

Currently available primarily on satellite tv providers, the ads are guaranteed to run in the household they are purchased for.

“It doesn’t matter what you’re watching,” Michael Palmer, president of i360, a conservative-leaning firm that offers the service, told the Washington Post.

The technology is still in its beginning stages and relatively expensive, but offers promise, as well as privacy concerns.

One thing that’s clear is that by the next election cycle, the old days of 1000 GRPs will be over.

“I think we’re getting very far away from that view,” said Kantar’s Wilner.

COMMENTS

Please let us know if you're having issues with commenting.