Pandora Is the Latest to Tap Political Ad Revenue, Won't Be the Last

On Friday, streaming music service Pandora announced that it will begin accepting political ads. The ads will target voters based on the ZIP code entered at registration.

This could not have been a hard decision for the company. Citizens United has ensured that an already money-soaked process will see bushels more donors looking to invest where they can. Earlier this month, the Times noted that the river of cash is reshaping local TV news, spurring consolidations intended to streamline cash intake before 2012 gets into full swing. In 2008, the McCain and Obama campaigns spent a combined $370 million on TV ads; that total will certainly go up next year.

The goal of a modern political campaign is simply articulated, if complex in practice: delivering a particular message to a particular voter. It's microtargeting: fracturing the pool of voters into pieces that can be pitched precisely. Nearly every financial decision is made with this in mind: how can we deliver the best possible message to the best possible voter at the lowest possible cost?

The process works as follows, at least for a race above a certain budget. A poll is commissioned, revealing existing bases of support and messages that resonate for particular populations. Talking about your candidate's work on the school board is popular with Asian women, say, while her opposition to gun control legislation appeals to white men in the suburbs. This knowledge, combined with public data about voters, lets campaigns do two things: try and increase the number of supporters who come out to vote, and convince likely voters to vote for their candidate.

Recommended Reading

For a campaign, money spent is money that has to be raised. So they balance considerations: cost, ability to target a message, size of potential audience, efficacy. Sending direct mail, for example, allows you to target individual voters, like Latinos in a particular zip code. It's not too expensive for a small audience. Broadcast television on the other hand lets you hit a huge swatch of people, but that reduces the targeting - and it's more expensive.

The chart below, which I constructed from personal experience and ran by some consultants, provides a rough overview of how techniques compare. Circle size is potential audience. The darker the blue, the more effective it is in communicating a message; I'll admit some subjectivity in that designation. (Pandora, Twitter and Facebook are new enough in this space that it's hard to judge efficacy, hence the shading.)

You can see that the ideal technology would be a giant, dark blue circle covering the upper left quadrant. It doesn't exist. What campaigns do, then, is some highly targeted, low-cost tools mixed with a broad blanket covering everyone else.


Today, the blanket is television. On broadcast channels, you can reach a huge audience. On cable, you can refine your targets (think: Golf Channel). You can pick a location -- helpful in races at every level. It's proven. People remember the ads they see. And, like lawn signs, candidates love them. They like to see their name out in the world.

But the peak moment of influence for TV ads is near. Broadcast's audience is softening; cable networks' targeting is unrefined. Both hit wide swaths of people who don't need to see the ad. City-level targeting used to be enough. Now, campaigns want to influence city blocks. They'd love to lose the blanket forever.

Which is why Pandora is in such a good position. The service already differentiates advertising by location; if you're a listener, you've no doubt heard location-specific ads pop up. You may be less aware of the way it subtly differentiates along another line: what you listen to.

An example. One day, I was listening to a station playing Mobb Deep, a hip-hop group from Queens. The ad on the page was for a dating service, and the background was a collage of women's photos, all African-American. I switched the station to A Tribe Called Quest, a group with more mainstream audience. When the page reloaded, the ad remained, but over a background collage of white women. The service is predicated on understanding the relationships between songs and the way people respond to them. There's no question that they can make accurate predictions about a slew of demographic traits as easily as they can figure out what song to play next.

What Pandora offers a campaign is zip code-level targeting (more refined than broadcast or even most cable) and breakdowns by demographic, for a price that will undoubtedly beat TV's. Its audience size is nowhere near a TV broadcast, of course, but it's closer to becoming that big blue circle in the upper left corner.

Every indication is that the trend will continue -- that advertising (and, therefore, political advertising) will grow increasingly able to discern the sort of message that will get us to buy based on who and where we are. Pandora's certainly not perfect; for example, it still thinks I live in New York City, offering deals that are not particularly useful to me. But Google, Facebook, Twitter and any number of other companies are competing to develop that precision. Someone will develop it. The carpet-bomb TV ads will be abandoned for the laser-guided sales pitch.

It's fair to be worried that clunky, clumsy TV won't be the only casualty. Why have a town hall when you can tailor you message more cheaply and safely through a Google ad? Why debate, when you can present opposition research on someone's Facebook wall? Why campaign when you can market?

Earlier this month, the Washington Post's Ezra Klein reviewed a paper which presented a new theory of the role of the modern political party. Political parties and their candidates, the theory goes, are no longer motivated to reflect the desires of the voters. Instead, they reflect the whims of a narrow set of stakeholders. Campaigns are not listening tours from which a platform is derived and evolves; they are sales pitches, in which voters asked to buy a candidate are told what they want to hear. It's not an election, really: it's brand dominance. 

Pandora's business isn't maintaining the sanctity of our political system. Streaming music and evaluating preferences is. Pandora presents something of potential value to political campaign customers, even if that value will inevitably be eclipsed by something narrower and deeper in focus. Opening the service up to political ads moves forward the sophistication of political targeting and makes them money -- potentially, lots of money.

This is a capitalist democracy. Who could say no to that?