Below are a few examples of my past work leading various cross-functional teams at Spotify. I have a wide range of experience building consumer products that millions of people use every day.  I believe the best products combine the right mix of user experience, data and content to create an emotional connection with users.


Prototypes of proxy search on the left and what ended up shipping on the right. This initiative was about making successful intents even more successful.

Proxy Search (Prototyping)

The search team had a number of initiatives to make failed intents more successful and successful intents even more successful. The concept of "proxy search" was developed for the latter. Take the example of Track searches which are a dead-end after one play and are the second most common search content. Proxy search could keep the music going here. But there were also several other interesting applications that the team prototyped during the design phase.

Results

I left before finishing testing of the concept.  The team later shipped one iteration of this (note the U2 example on the right).


Sell your merch on Spotify

Presenting merch on Spotify to artist's fans was the first component Artists had full control over inside their artist page on the service. Whether you're a Lady Gaga fan and want your "lady fucking gaga" shirt or need your vinyl fix for Arcade Fire's latest album... you can now browse all this unique content on the service (purchases made through Topspin).

Results

Launching Merch was a massive step forward for the company toward developing significant mindshare within the music industry as a promotional vehicle (building on what we did for concerts). This was measured primarily qualitatively at the time. However, the number of artist sign-ups for merch within the first 2 weeks (1,000) was phenomenal and surpassed our expectations. This feature lives on today.

Artists have full control over what merch they want to sell through Spotify. 


Concert discovery on Spotify reduced the need for leaving the service and surfaced the information at the moment when you're most interested. it's become such a popular use case on the service that Spotify continues to build on this concept today.

Concerts Discovery - Spotify

Spotify users can see when their favorite artists are playing shows in their areas, browse more information about the shows, and click to get their tickets. These new concerts, originally powered by Songkick, also surfaced in the Discover feed for artists you follow.

Results

Within the first year, Spotify was driving between $2.5m - $5m in concerts sales for our artists. This has only increased over time. The success of this execution was in its win-win approach for both users and artists. The value of Spotify as a promotional tool started at this point. And, this feature lives on today.


Discover - Spotify

Discover was Spotify's first attempt at imagining what a personalized music experience could be. The lean-back discovery "magazine" concept tested as Spotify's home page for a brief period. Users listen and follow music as signals for Discover to recommend new content such as album reviews by Rolling Stone, playlists, concerts, artists, and tastemakers. The picture to the right is from Daniel Ek's live demo in New York which I proudly attended back stage and helped rehearse. For the demo we used our new Web desktop version of Spotify to show off the magazine look & feel. Daniel followed a few of his favorite artists and at the end got a notification on his iPhone that one of them (Bruno Mars) had just released a new single.

Results

Testing indicated that the global maximum potential for this product was minimal since the overall positive retention effects were within the margin of error. So, the team optimized for the local maximum (engagement, follows, etc.) and moved on. However, the same discover algorithm ended up being a core part of the company's lean back playlist strategy. It's called "Discover Weekly".

Daniel Ek, CEO of Spotify, on stage unveiling Discover for the press in December 2012. The screen shown here is about the size of a 30ft TV. Discover served up personalized recommendations based on listening and follow data.


Remember seeing lots and lots of music in your Facebook feed from what your friends were listening to? That was us. And it worked really well.

Spotify's deep integration with FB Open Graph

Our deep integration with open graph using tokenization was cutting edge and rock solid thanks to our amazing engineering team. A user clicking on a Spotify play button downloaded the app (if they didn't have it) in the background, logged the user in using their FB credentials (w/ token) and started playing music. All of this happened so fast that the user didn't need to think about it and the music was free!

Results

This was a tremendous growth success increasing MAU 200% (10MM to 20MM) within the first few months of launching (from flat line growth).