How Spotify Uses Analytics To Deliver Its UX

The streaming giant is smart in its use of data


Given how popular Spotify is and how familiar most people are with it, it’s surprising to think that, before its launch in 2008, users still predominantly paid for music downloads. As the primary software to go with its ubiquitous iPod, Apple held the keys to the music industry and paying $0.99 for a track was the norm. Fast forward nine years and Spotify is second only to SoundCloud in terms of user numbers (some 100 million). It’s $2 billion revenues are still not been enough for the company to be profitable, but its safe to say it understands how users want to handle their music collections.

Even when you ignore its intuitive user interface and its seemingly endless supply of music (30 million tracks, roughly), Spotify delivers a unique experience. How? It’s powered by data.

In 2012, Spotify launched its ‘Discover’ tab, which included primarily new music from a user’s favourite artists, but also introduced the idea of recommended artists based on listening history - it’s part of the company’s ultimate success, much in the same way that Netflix has grown off the back of its own sophisticated recommendations system.

Later, Spotify added recommended tracks at the end of playlists for users to add based on the playlist’s content, and Discover has developed into a weekly playlist of fresh music designed to expand users’ musical horizons based on their previous listening history. In 2015, the streaming service acquired Seed Scientific, a data analytics startup that had previously worked with Beats Music (now absorbed into Apple Music).

The unit became part of Spotify’s data analytics unit, along with The Echo Nest the following year, a music data specialist startup that analyzed over 35 million songs. From this analysis, a trillion data points were mined to help Spotify make better recommendations. Today, Spotify offers its users sophisticated recommended playlists curated algorithmically, including both the users owned music and music the user may be unaware of. The ‘Daily Mixes’ are remarkably reliable, such is the depth of Spotify’s user data.

This data is extremely powerful, and Spotify also uses it to shape its marketing campaigns the world over. If a particular marketing campaign worked well in Chicago, it could run a similar campaign in Den Haag and expect a positive result. Why? Because on music preferences, the Netherlands’ Den Haag is more similar to Chicago than Melbourne, Texas, Berlin or even New York. It’s this level of insight that the streaming service simply couldn’t get without the use of advanced analytics.

It also uses data to expose users’ guilty pleasures in its targeted marketing efforts. One particularly tongue in cheek campaign ran in the very trendy, very sophisticated neighbourhood of Williamsburg in Brooklyn. The message: ‘Sorry, Not Sorry Williamsburg, Bieber’s hit trended highest in this zip code’ was displayed on a huge street advertisement, a piece of marketing made engaging and funny only by virtue of its localization.

Spotify also has in-depth data on users’ listening habits depending on different weather conditions and is able to plan its weather-based media buy with this in mind. In a hyper competitive market, Spotify is a shining example of a company using its data to stay ahead of fierce and (often) better funded competition. Making smart decisions around recommendation services and not being afraid to be tongue-in-cheek with its marketing, the music streaming service has been able to report steady growths in revenue, and an improved bottom line looks inevitable.

Vision small

Read next:

Big Data Forecasting In Pharma