Apps & Mobile

Pandora’s Biggest Weakness, Potential Opportunity


By Kyle Bylin (@sidewinderfm), founder and editor of, a music and tech think tank.

Pandora has evolved during its tumultuous, decade long rise to ascendancy in the
Internet radio sector, but the product is not perfect. Avid users lament about the repetitious nature of their custom stations and their tendency to pander
toward certain artists. The alternative rock band Weezer, for example, is a chameleon whose wide ranging and popular songs eventually squeeze their way into any station.
Furthermore, custom stations that only play songs you like from artists you love can become a remote island in an increasingly social world.

For all of the
joy that users derive from escaping the clutches of traditional radio where they lacked a say in what gets played and suffered through endless car commercials, they listened to and discovered music alongside other people. This shared experience either bound them together or culturally set them
apart. The same cannot be said of the social features that Pandora offers. Its social integration with Facebook and Twitter allows users to share favorite songs and stations outward, but aside from user profiles and music feeds they are isolated within. 

These several weaknesses, among many others, are explored by four influential product executives in the music and technology industry, who submitted answers to's panel question:
What is Pandora’s biggest weak spot in the service? If you were put in charge, how would you improve on the product?

Pandora Should Bridge Offline and Online Experiences


Max Engel is the Director of Product at the web publisher SpinMedia, formerly BuzzMedia.

From a discovery standpoint, Pandora is built upon the concept of reinforcing existing taste and presenting a variation on a theme. In practice, this means
that a listener is exposed to music they’ll like, but not necessarily new music. Pandora could add a means by which it scans a users music library to
determine existing artists already in their collection. Then, truly “new” discovery could occur, and a listener could stumble across new bands and purchase
new music.

Additionally, Pandora is inherently a passive music experience, and so there is little engagement with the product outside of the minimal amount of
interaction necessary when listening to a station. While they’ve dabbled with social elements through the website, it would be interesting for them to
focus on other threads of music discovery. For example, Bandsintown lets me link my taste profile to Pandora to help me explore local concerts. I think
Pandora should more deeply explore ways to bridge offline and online music experiences.

Finally, Pandora should explore the ways in which they could act as an intermediary platform to connect labels and artists with listeners. For example,
Pandora is sitting on a trove of valuable analytics that get generated as a byproduct of user interaction. This data could be immensely valuable if
packaged correctly, and the music industry would be willing to pay for access to a service that would help them better understand their consumers.
Similarly, there has been a great deal of activity in the direct-to-consumer music marketing space, and Pandora could also explore how they could offer
better services in this area to allow a band to better engage with their fans.

Pandora Should Enhance Social Features, Add More Context 


Vivek Agrawal is the co-founder of the playlist service Playground.FM.

First and foremost, I want to acknowledge that Pandora has done an amazing job building their product and establishing mindshare in Internet radio. Tim
Westergren, Tom Conrad, and the rest of the team have spent years creating such a beloved product and should be praised for their success.

I think the biggest weak spot in the service is that in an era where music has become such a personal experience we've begun to lose the emotional element.
Music used to create a dialogue amongst friends and music fans — scanning through CD and MP3 collections, creating and swapping your favorite mix tapes,
even occasionally listening to music together. The personalized radio has unfortunately removed people from a listening experience which used to be shared
with others.

In Pandora, I feel like I'm escaping into my own music world, but I feel alone. In contrast, social music services like Spotify,, and Rdio
have me exploring a world of music through other people; when I find a great song, I want more music from that person, and more interestingly, I want to
know more about that person — it's a rare chance to connect through music. Even terrestrial radio's talk shows and DJs give a human touch to the listening
experience, that I'm not just listening to music for music's sake, but that I'm getting into someone else's personality through their music.

Improving on someone else's product is always hard since I would never have the full visibility into what would make the product "better" for its userbase.
But I think the key is to bring social elements into Pandora without compromising their genome or listening experience. With that, I would look into ways
that Pandora could expose the people that actually help power the genome. The musicologists and curators that Pandora has recruited to build our favorite
stations are probably fascinating people, and I would love to learn more about them and their personality. Maybe they could provide individual notes around
why they like the song and recommend other songs they also like.

Pandora Should Expand Library, Broaden Recommendations


Tony Hymes is the community manager at the music
discovery website whyd.

The biggest issue with Pandora is its repetitiveness. Tim Westergren, one of Pandora’s founders, recently stated that they have over one million tracks
from 100,000 artists. That might seem like a lot to some but is a drop in the bucket compared to Spotify’s catalogue of 20 million tracks, and doesn’t even
register with the amount of music available on YouTube.

Since Pandora’s library is relatively very small, similar artist stations can be exactly the same. Try listening to The Temptations radio then to Diana
Ross and the Supremes radio and then to Marvin Gaye. There will be almost continuous overlap, as if the songs are coming from a Motown grouping and not
individual artists.

This problem is only magnified over the years that I’ve used Pandora. It is now pointless for me to listen to any of the above-mentioned stations, since I
know exactly what I am going to hear. I don’t know how well Pandora does with more modern music since I don’t use Pandora very much any more, but my hope
is that they would continue to add new music and help to classify that music into radio stations that might have more traditional music (if the genome
project matches it correctly, of course).

One negative result is that there is not a lot of cross-decade recommendations, the music seems to be limited by period; I would try to add more types of
new music that get closer to traditional styles, such as adding some remixes of well known oldies in a Stevie Wonder radio, or even some nu-disco to Martha
and the Vandellas radio. Maybe the music genome has too many criteria for matching (4,000 data points per song) which eliminates the spontaneity and close
approximation of music taste that might be more, ahem, “human.”

Pandora Should Help Users Decide What To Listen To Next


Jason Keck is the founder and CEO of the social music application Stereotypes.

The biggest weak spot in Pandora’s service is something that was highlighted with the launch and success of Songza. Before I discovered Songza, I had no
idea that I had trouble deciding what to listen to. Now that I have Songza, I feel like picking a station on Pandora is so hard! Especially when I’m
looking for something fresh and have no idea where to start.

The challenge for Pandora, as with any consumer app today, is to keep the user experience simple, consistent, unique and memorable while also being forward
thinking in technology and user experience. Pandora’s core, memorable experience is based on three simple things today: 1) a user selecting an artist as
the seed for a station, 2) the ability for a user to give feedback to make sure the station plays music they like and 3) an underlying technology that
plays similar songs based on the seed song and other “liked” songs on the station. I would argue that two and three are deeply ingrained in a user’s
understanding and trust in the brand. Selecting an artist as the seed for a station was novel 10 years ago, but times have changed.

Pandora has added “genre” stations to help users get started, but even that is an antiquated approach. Adding the ability to define the mood I want or
activity I’m doing to help me pick an artist station would really help me get started. That would also opens the door for some very interesting social
features. I’d love to know what my brother listens to when he works out, or what my wife listens to at work. Or to pick a mood with genre and then see then
see a list of stations created by my friends that match. is founded and edited by Kyle Bylin of Live Nation Labs. If you would like to contribute a post to be featured on the site, please reach out.

Share on: