MarketingData-Driven MarketingSpotlight On: Pandora’s Jack Krawczyk

Spotlight On: Pandora's Jack Krawczyk

The director of product management recently spoke with ClickZ about Pandora's troves of user data, which goes back more than eight years, and how it enhances the music-streaming platform.

If there’s one thing Pandora has in spades, it’s songs. But if there’s one other thing the music streaming platform has in spades, it’s data on more than 250 million users, none of whom should have the same experiences, according to Jack Krawczyk, the director of product management at Pandora.

ClickZ recently spoke with Krawczyk about how data helps Pandora, some of the platform’s targeting plans, and what he’s listening to right now.

pandora-jack-krawczykClickZ (CZ): Pandora has data going back more than eight years. How do you turn this vast amount of big data into valuable data?

Jack Krawczyk (JK): Let’s get geeky on this one. As we build up our data set, the integrity of the predictions we make in our advertising algorithms grows stronger. Data is our oxygen. Using standard machine learning tactics, we can use the relationship between disparate data sets in multiple dimensions to predict things like, people who listen to jazz more than any other genre tend to have traits that indicate they’re likely to vote Democrat, which allows us to better personalize the advertising experience for our listeners and our partners.

CZ: In October, you opened up data to artists. Have you found that they’re surprised by the things they’ve found out about listeners?

JK: Whether it’s artists or advertisers, it’s always fun to see the reaction of our partners to the insights that we can generate. Sometimes we even get surprised. For example, we had a pharmaceutical advertiser develop a brand station that was creating a soundtrack targeting women going through menopause. When we launched, we knew it would be a risk, though on a per-listen basis, the station (targeted to females 45-64) turned out to be one of the most engaged genre stations we produced. It just goes to show, when music can be presented in the right context and drive people toward feeling like they’re associated with a community that is going through similar moments in life, it can resonate with anyone.

CZ: You’ve drilled into user data in the past to create targeted audience segments including the Hispanic and Spanish-speaking community. How did the initial targeted ads go? Was there a drastic lift in engagement?

JK: The targeting initiatives based on Pandora’s prediction algorithms continue to drive meaningful performance for our advertisers, and a more personalized listening experience for our listeners. While engagement rates rise, we find that with richer targeting in place, we are able to drive longer listening sessions from our audience. As we look to 2015, expect to see more options for advertisers to carve out their target audiences.

CZ: What’s your top Pandora station?

JK: Mark Ronson Radio is getting the most spins from my account this month, though I must admit I thumbed down “Uptown Funk.”

CZ: A few months back, Spotify teamed up with Uber, and Shazam is working with a company that does digital watermarks. How do you think platforms’ user data can enhance these partnerships?

JK: The power of the platform is the key to winning the attention of consumers. Our goal with Pandora is to build a service that is entirely device agnostic, allowing the consumer to determine where and how they want to listen to Pandora. We most recently launched our native Xbox One app in December and have been pretty pleased with the initial results.

CZ: And lastly, with the vast amount of data out there, do you think the plot of Live Free or Die Hard could ever really happen?

JK: I can’t recall the plot of Live Free or Die Hard, so it’s hard to comment on that.

[Editor’s note: If you have seen this movie – in which a data scientist hacks into government and commercial computers with a goal of crashing the U.S. financial system with a fire sale of financial assets, before Bruce Willis saves the day, obviously – and can answer this question, please do so in the comments. We’re curious.]

Related Articles

5 businesses using AI to predict the future and profit

AI 5 businesses using AI to predict the future and profit

1y Clark Boyd
How can analytics and AI allow marketers to predict the future?

AI How can analytics and AI allow marketers to predict the future?

1y Clark Boyd
How deep learning is changing the game for both advertisers and consumers

Data-Driven Marketing How deep learning is changing the game for both advertisers and consumers

1y Daniel Surmacz
Tencent's Emily Ma on the challenges for marketers in China

Analytics Tencent's Emily Ma on the challenges for marketers in China

2y Sophie Loras
What are the biggest obstacles to programmatic advertising?

Data-Driven Marketing What are the biggest obstacles to programmatic advertising?

2y Rebecca Sentance
Three ways to create insights from consumers’ click histories

Actionable Analysis Three ways to create insights from consumers’ click histories

2y Deren Baker
Keeping the right balance of "person" in "personalization"

Data-Driven Marketing Keeping the right balance of "person" in "personalization"

2y Mike O'Brien
How does Pandora know its users so well?

Data-Driven Marketing How does Pandora know its users so well?

2y Mike O'Brien