Jump to Any Race

Cook House Popular Vote Tracker

  • Democrats 34,927,573 (45.48%) | Republicans 39,651,654 (51.64%) | Other 2,211,156 (2.88%)
Political Advertising|By Elizabeth Wilner, October 22, 2013

GRPs aren’t the end all, be all anymore. Elizabeth Wilner’s biweekly column is now a weekly compilation of news and analysis covering the fast-growing junction of advertising, Big Data and politics. Thank you in advance for your tips, comments and gentle critiques: elizabeth.wilner@kantar.com.

FIRST IN ‘ON POINTS’: KANTAR MEDIA CMAG’S $2.4B OVER/UNDER FOR 2014 TV AD SPEND. Using CMAG data and past Cook Political Report  ratings, the University of San Francisco’s Ken Goldstein puts the over/under on local broadcast TV advertising for the 2014 elections at $2.4 billion. A drop in the number of competitive races overall, and in the number of top-50 markets hosting competitive races, accounts for what may amount to a very modest increase in TV ad spending from 2010.

Here’s the math: Goldstein expects around $6 billion to be spent on the midterms, with about half that amount going toward TV advertising. Of that $3 billion, Goldstein estimates that just under 80% will go to local broadcast for a total of $2.4 billion—an increase of just $100 million from CMAG’s ultimate 2010 total. Spending could range higher or lower than that estimate if there’s any variance in the amount raised, the proportion spent on TV, or the distribution between broadcast and cable.

Factoring heavily into Goldstein’s calculations are the anticipated number and location of competitive races in 2014 compared to previous cycles. While the elections are more than a year away, at this point, Cook sees far fewer competitive races at the federal and gubernatorial level. Fewer competitive races means fewer ad dollars—and especially fewer ad dollars paid by outside groups that are not eligible for the Lowest Unit Rate stations are required to provide candidates within a 60-day window of Election Day.

This table looks at the number of Cook-rated competitive races (i.e., either Lean or Toss Up) in the two prior midterm election cycles of 2006 and 2010, as well as the current outlook for competitive races in 2014. The first number is the total number of competitive races; the number in parentheses is the number of races rated as Toss Up. Again with the caveat that things can change, Cook sees nine fewer Toss Ups for Senate, 34 fewer Toss Ups for the House, and 12 fewer Toss Ups for governor in 2014 than at this same point in 2010.


2006 2010 2014
Senate 13 (8) 17 (11) 11 (2)
House 74 (28) 84 (46) 43 (12)
Governors 13 (6) 22 (18) 11 (6)

The drop in the number of competitive statewide races is significant given their outsized role in driving TV ad spending, but the total amount of local spot spending is also influenced by where competitive races are located. Fewer US media markets seem likely to attract significant TV ad spending in 2014.

In 2010, for example, 33 of the top 50 media markets hosted at least one competitive statewide race. Eight markets hosted hot races for both Senate and governor. Four of those eight markets were located in California and saw pricey primaries in addition to general election contests. While high-profile referenda always get TV ad dollars flowing in California, the state has no Senate contest next year and, at least at this point, not much of one for governor.

Compared to 33 of 50 top media markets in 2010, only 12 of the top 50 markets are hosting a Senate or governor’s race with a present Cook rating of Lean or Toss Up. Not only are there fewer competitive statewide contests in 2014 compared to 2010, but the air wars will take place over Little Rock and Anchorage, not Los Angeles and San Francisco.

There’s also a clear correlation between the number of spots aired (i.e., not the number of unique ads, but the number of airings of those ads) and the competitiveness of a race. In Colorado in 2010, both parties hosted competitive Senate primaries plus a hard-fought general election. Cook rated the race a Toss Up from the start. Ohio’s Senate race, on the other hand, started off looking very competitive but seemed safer for Republicans by the fall.

Since states differ in advertising expense, finding the right apples-to-apples metric is tricky, but Goldstein used the two states’ largest media markets: Denver and Cleveland. Conveniently, they’re the 17th and 18th largest media markets in the country. Denver saw 27,000 spots, of which more than half were aired by advertisers who didn’t qualify for the Lowest Unit Rate. Cleveland stations saw just over 4,000 spots, with only about 10% sponsored by groups.

Having governor races in one-third of the United States somewhat compensates for the lack of a presidential contest at the top of the ticket—but only somewhat. Even in governor races with contested primaries, the ad spending typically doesn’t reach the level of a contested presidential primary. And while a state’s airwaves can be loaded with ads for governor (as in Virginia today) in October, the ad spending doesn’t tend to begin as early or remain as sustained as it does during a presidential race.

The silver lining for stations, Goldstein says, is that TV advertising rates were lower in the post-recession midterm cycle of 2010 and have rebounded since. And while outside groups may have fewer competitive races to try to influence than they had in 2010, they probably will account for a greater share of the advertising activity that occurs. We may see fewer spots but just as many ad dollars.

Coming next week: “On Points” looks at developments and expectations for 2014 local cable ad spending.


WHAT IF WE SEE A CONTESTED PRIMARY EPIDEMIC? With contested primaries meaning springtime air wars, we’re keeping a close eye out for Republican nomination battles—or Republican primary challengers who may become more viable—due to intraparty fallout from the government shutdown and debt ceiling fight.

Cook Political Report Senate Editor Jennifer Duffy points out, “A majority of GOP Senate incumbents already have Tea Party challenges, Matt Bevin's challenge to Mitch McConnell being the most serious so far.” Challenges are brewing against John Cornyn of Texas (primary in Q1), McConnell in Kentucky, Thad Cochran in Mississippi and Lindsey Graham in South Carolina (all in Q2), and Pat Roberts in Kansas, Lamar Alexander in Tennessee, and Mike Enzi in Wyoming (all in Q3). (Cochran has not yet said whether he will seek re-election.)

Alexander, Cochran, McConnell and Graham voted for the bill to re-open the government and raise the debt ceiling; Cornyn, Enzi and Roberts voted against it. “Most of these challenges are pretty minor right now—the question is whether there’s enough significant fallout from last week's vote to give any of these challengers momentum they don't currently have," Duffy advises.  She adds that all seven incumbents will come under pressure from their opponents and Tea Party groups when the Senate revisits these issues early next year. In fact, the pressure already has started in Mississippi, where the Club for Growth has a new TV ad boosting Cochran’s primary challenger.

On the House side, Editor David Wasserman has his eye on brewing conservative challenges to Reps. Pete Sessions (TX-32), who’s up for renomination in Q1, and Walter Jones (NC-03), Mike Simpson (ID-02) and Bill Shuster (PA-09), all with primaries in Q2.

ON(LINE) POLLING: WHEN AN ONLINE POLL IS A “POLL”… In a quiet but tectonic shift, two national news brands have undertaken online public opinion research this month. Most significantly, the Associated Press became the first US national news brand that conducts its own regular public opinion polling to switch from telephone to online interviews. Per director of polling Jennifer Agiesta, the AP will now conduct all its regular polling through GfK’s “KnowledgePanel,” though they will use GfK’s telephone omnibus survey as needed.

Agiesta explains in her column, “The science behind polling says that in order for survey research to work, respondents must be randomly selected rather than self-selected and everyone in the population must have a chance of being asked to participate. GfK addresses this by recruiting its [KnowledgePanel] from a random selection of all Americans with either a telephone or a home,” then providing a connection and a laptop to participants who don’t already have access.

ON(LINE) POLLING: …AND WHEN AN ONLINE POLL IS A “SURVEY”
. On October 15, NBC News, in partnership with Esquire, released an online poll mapping out America’s “new political center.” The survey (as NBC is calling it to distinguish it from its regular NBC News/Wall Street Journal telephone poll) was conducted by the bipartisan team of the Benenson Strategy Group (D) (disclosure: owned by Kantar, my parent company) and Public Opinion Strategies (R).

We found the fact of the online survey and the methodology nearly as interesting as the results. According to Danny Franklin, who oversaw the poll for BSG, the methodology had been developed by the firm for political ad testing for various clients (though not the Obama campaign). This survey marked the first time BSG has used it for a news media poll. “We believe that the data themselves are more faithful to people’s true opinions because they’re able to think about and reread questions before they answer, enabling us to ask a more complex and, we think, useful type of question,” Franklin says of the 140-question poll. Some of the questions were complicated enough that they simply could never be asked over the phone, while the sliding scale of 1-9 gave respondents the chance to really register the intensity and nuances of their views.

Using a system called VoterInsider, a joint effort of BSG’s sister company iModerate and Catalist, all survey participants were matched to voter files to confirm that they are indeed registered voters. This enabled BSG to avoid the primary concern with online polling: that you can’t be sure the participant is a voter. 

To comment on anything you’ve read here: elizabeth.wilner@kantar.com.