Spotify’s Trend Research with Social Listening

Spotify’s Trend Research with Social Listening
28. March 2018 Nina Walloch

Spotify’s Trend Research with Social Listening

Music streaming is dominating the modern music business. Providers such as Spotify have to know about recent trends and the preferences of their users and set up various playlists with relevant and new songs. Spotify’s success is significantly impacted by their data-driven approach and the use of trend research. With this article, we would like to emphasise why trend research with social listening plays such a fundamental role for music streaming providers.

spotify trend

Changes in the music industry – From manual curation to Big Data

Music streaming is one of the most prevalent formats in the modern industry with continuous global growth. 112 million people worldwide are now using paid subscriptions for music streaming (ifpi) Рand this number is likely to grow even further. Music streaming providers respond to the changing listening habits of consumers who are increasingly in favour of playlists with a good mix of different songs, artists and genres. They generally prefer this to listening to whole albums of an artist at once.

Nowadays, being featured in playlists determines the success of artists. 

Examples of famous artists like Katy Perry or Ed Sheeran show how important it is to be featured in many playlists so that their new songs or albums are quickly discovered. Spotify generated a lot of attention for the new singles of both artists who then reached a new streaming record after a very short period of time (The Verge). It is quite obvious that the success of modern artists does not solely depend on album sales anymore but, perhaps, more so on their streaming time.

Spotify, currently the market leader in the music streaming industry (ifpi), relies heavily on data and trend research with social listening for personal playlists and a better user experience.

The revolutionary recommendation model of Spotify

Spotify has made considerable investments in Artificial Intelligence in order to be able to improve music discovery and distribution with music curation features such as the Discover Weekly or Release Radar playlists. Even though most of Spotify’s 2 billion playlists are made by the users themselves, their own curated playlists are attracting millions of followers (Economist). The powerful source of this success is Spotify’s recommendation algorithms that are based on a mixture of best practice strategies of the industry.

Spotify’s recommendation model is based on three approaches:

  1. Collaborative filtering
    Spotify’s user data contains a lot of information about listener behaviour, for example what tracks are listened to or which songs are saved to personal playlists. As these insights are collected for many users, Spotify can make automatic predictions about the musical interest of a user and recommend songs and artists to him/her Рbased on similarities with other users.
  2. Audio analysis
    In this approach, Spotify analyses key characteristics of songs to find similarities between them. Based on a user’s listening history, Spotify will recommend similar songs which they might enjoy as well.
  3. Trend research with Natural Language Processing (NLP)
    By employing NLP models, Spotify can track online publications and conversations, such as articles, blogs or social media posts, to identify trends and fulfill the needs of their users.

The huge potential of trend research for Spotify

Algorithms usually make recommendations mainly based on previous behaviour only. Performing trend research is therefore highly necessary in order to be able to recommend new music and new artists to the users. By crawling the web for relevant trends, artists and topics being discussed online, Spotify can figure out what people are saying about artists and songs, the sentiment of these conversations and which other artists are mentioned at the same time.

With the help of trend research, Spotify can make content-based recommendations that make playlists even more varied and more personal.

Spotify regularly suggests new music in its discovery features such as Discover Weekly and puts together other playlists for specific times of the day, activities and moods Рall personalised to users’ music preferences. The company seems to have found the secret ingredients for playlists that feel refreshingly new and personal at the same time.

Playlists like Discover Weekly seem to convey the feeling that there is a real person behind the playlist that knows and understands you and your taste in music. This personal connection to the product is a very effective way of retaining customers and convince them to choose Spotify’s streaming service over others; giving them a competitive advantage.

As social media and the user listening experience seem to be increasingly connected, trend research with the help of social listening is necessary in order to realise these advantages in the music streaming business.

Market and trend research with social listening

With the help of social listening and intelligence tools, streaming companies, such as Spotify, can continuously monitor public communication streams in order to discover conversations, opinions and thoughts of their target group. Social listening tools, such as Ubermetrics Delta, also track conversations on very specific online channels that companies did not necessarily know existed; such as niche music blogs or discussion boards. On these publications, music lovers share their insights and, by keeping track of these opinions, Spotify can discover fairly unknown and new artists for their discovery playlists.

With additional features such as the sentiment analysis in Ubermetrics Delta, they can further analyse consumer mood, mindset and values to incorporate these attributes to their playlists. Spotify has put together various playlists for specific moods and situations with songs that represent the desired emotional state without interruptions by unsuitable songs.

Spotify

Marketers can also use trend research to create real-time marketing campaigns. Spotify has picked up this opportunity in the past. They have used their insights to showcase the listening activity of its users on outdoor billboards:

On the day of the Brexit vote in the UK, for example, billboards displayed:
‚ÄúDear 3,749 people who streamed ‚ÄėIt’s The End Of The World As We Know It‚Äô the day of the Brexit Vote. Hang in There‚ÄĚ

On Valentines Day, they were showing:
‚ÄúDear person who played ‚ÄúSorry‚ÄĚ 42 times on Valentine‚Äôs Day – What did you do?‚ÄĚ

By picking up relevant and widely discussed topics, global events and trends, and combining these with personal insights such as the listening behaviour, the campaign is perceived as relevant and at the same time very relatable by Spotify users.

Conclusion:

Spotify successfully uses trend research with NLP models and social listening in order to continuously extend and improve their playlists and make them more personal. Trend research enables companies to keep track of recent topics and trends and to profit from these valuable insights in their daily marketing and PR efforts.

Junior Marketing Manager