How Pandora Aims to Double Its Number of Relevant Ads With Cross-Device Data

Music service hones targeting with Krux

Headshot of Christopher Heine

Pandora last week said its third-quarter ad revenues rose 31 percent year over year to $255 million, seemingly underscoring the narrative that the digital music service attracts brands because of its engaged listening audience.

But in an increasingly fragmented media world, the Oakland, Calif.-based company—like so many tech platforms—strives to better target consumers across desktop and mobile screens. Not to mention, go toe-to-toe with for marketers' attention with Spotify and Apple Music.

So, it's been working with Krux, a data management platform that aims to increase the relevancy of the promos Pandora serves to its 78 million users. Pandora offers audio, video, display and native ads. 

"Eighty percent of Pandora's listening occurs in the Pandora app on mobile devices," said Dave Smith, vp of monetization and yield at Pandora. "However, most of the data in the ad ecosystem today is keyed by cookie. To deploy that data to make ads more relevant, we have to find a way to bridge the data from the cookie world to the in-app world.

"The good news is that Pandora has a large number of listeners who we see logged in in both environments. Leveraging that, we built a solution a few years back that is allowing us to match between 8 percent and 30 percent of listeners to advertiser and third-party data sources. But that's just not good enough. Our goal is to ensure that all of the advertising we deliver is relevant to the listener."

Here are a few examples of how Smith's team is working on that goal (be warned that this stuff gets pretty techie):

1. They marry Pandora users' anonymous ID numbers with IDs for iOS and Android devices as well as Krux's "master ID" system that tracks users across the Web but not including the walled gardens, so to speak, of Facebook, Google, Twitter, etc. This allows Pandora to see segments of consumers who share similar attributes—from being a hip-hop music fan to an avid Victoria's Secret shopper—from their behaviors across most of the Web. The behavioral data could include details gathered during Pandora registration, listening behaviors, on-site or in-app activity, third-party data providers and so on. To be clear, Pandora manages its own first-party data and only uses Krux to onboard second-party and third-party data.  

2. Pandora examines user profiles to build its segments for advertising clients. Segments can be as simple as "all female Pandora listeners" or as specific as "women in California aged between 18 and 24 who have listened to country music more than three times a week." So the data signals include behavior, interest, recency and frequency.

3. Such targeting parameters are stored in Krux's system. Therefore, when an ad campaign is created on the ad-tech player's server, targeted to a particular audience segment, ad impressions can be served any time the ad served recognizes users associated with that segment. 

"There is virtually no limit to the new and innovative techniques [Pandora] can develop to enhance audience experience and deepen consumer engagement," said Jon Suarez-Davis, chief marketing and strategy officer at Krux, which competes with other data-management platforms such as BlueKai, MediaMath and DataXu. 

How well the souped-up system will work remains to be seen—test results are not yet available. Smith of Pandora said the targeting system has attracted interest from Kellogg and ConAgra so far, as well as undisclosed financial and entertainment brands. 

"We are able to match more second-party [advertiser] and third-party data [other Web properties] against our listener base for targeting," Smith said. He added that the "match rate will be well over 50 percent, potentially doubling the number of relevant ads our listeners receive and doubling the revenue we can derive from those targeted campaigns."

Pandora's fourth-quarter earnings are expected to be released in January, so perhaps we'll see some of the fruits of his ad team's labors by then. 

@Chris_Heine Christopher Heine is a New York-based editor and writer.