Pandora (NYSE: P) will roll out a new advertising service next week that will allow political organizations to better target its listeners this fall. The targeting relies on location and listening data to predict political affiliation, and is part of Pandora's latest efforts to improve ad targeting, thus increasing the value of its advertisements.
Pandora increased its ad load last fall from four ads per hour to six ads per hour, and doesn't plan on increasing the load again in the near term. Instead, management plans on increasing ad prices through targeting. But targeting advertisements is much harder for Pandora than it is for a company like Facebook (NASDAQ:FB) or Google (NASDAQ:GOOGL), whose users volunteer much more data.
Learning about its listeners
Pandora's engineers are mighty clever. Even though Pandora listeners don't volunteer much more than their gender and zip code, the company is able to deduce a lot of information from its users listening habits. For example, a listener in her late 20s or 30s with a children's music playlist is probably a parent.
With political targeting, Pandora has taken it a bit further. By gathering voting data from each county and correlating it with the music that's popular in each county, Pandora can find Republicans and Democrats no matter where they're located.
Using independent data points requires a lot of guesswork, especially when there's not an exact connection between them. Therefore, the accuracy of the ads is not ideal, but Pandora's management believes its targeting system is 75% to 80% accurate. Comparatively, it's a safe assumption that about 100% of people that are members of the group "Young Democrats" on Facebook are indeed Democrats.
That's not to say Pandora's targeting is worthless. Indeed, there are numerous opportunities for the company to improve its targeting by using data from listening habits to infer interests. The problem is no matter what Pandora does, it will require a significant amount of guesswork. Maybe that hypothetical listener mentioned earlier is a preschool teacher and isn't interested in buying diapers.
As a result, advertisers aren't going to pay as much for Pandora's best guess compared to Facebook's no-brainer because the return on investment needs to match the market. If 20% to 25% of political ads miss their targets, they're practically useless to Pandora's advertising partner.
Do people pay attention to ads?
Every 20 minutes or so, a listener must sit through ads and listen to them. Pandora has implemented several methods in the last year to ensure that listeners are actually engaged with the music it's paying for. On the flip side, this offers some assurance to its advertisers that their ads are being listened to, but it's not a 100% guarantee. On a platform like Facebook or Google, advertisers only pay when a user engages with the ad.
Unlike with online video on YouTube, where it's almost a sure bet a user is sitting in front of her computer, music listeners are often multi-tasking. That's why YouTube and other online video companies are able to charge such a premium for their advertisements. Pandora, again carries a certain amount of risk that a user will not be paying much attention.
Pandora will begin rolling out ads in cars this year, however, which may garner a premium over its typical ads precisely because advertisers know that listeners will pay more attention. Pandora has more than 4 million in-car listeners, and it's firmly establishing its presence in many of the best-selling cars. The downside is that listeners are unable to take immediate action on advertisements like they are on desktop or mobile devices.
Increase ad rates faster than royalties
Pandora's royalty rate is set to climb 8% in 2014, and another 8% in 2015 before its agreement expires. The company is currently locked in a court battle that could determine the future of its royalty rates, but the increasing number of music-streaming services could potentially cause the rights to music to increase. In the near term, Pandora should have no problem growing ad revenue and improving its targeting. But there's certainly a limit to its targeting data, which means its ads will always demand less of a premium than more data-driven platforms.