Big Data: Spotify

Group 6

by: Buse Dikici, Jai Khanna, Jasmin Stallzus, Noah Wolford

spotify-logo-primary-horizontal-light-background-rgb

Spotify is one of the most known and used music streaming services in the world. Starting in 2008, it has now over 100 million active users including 30 million users that are paying for it. Spotify has over 20 million songs and everyday over 20.000 new songs are added to the database. Big data helps Spotify to use techniques and tools to improve the wishes of their users and propose recommendations. Without Big data Spotify would never have been popular as it is used nowadays around the globe. The key activities of Spotify is to maintain the platform because it is the only way of connecting with the customers. Any distribution of the service would result in customer loss, which would damage the company itself.

Background:

Spotify is a music streaming service which was launched in 2008 by Daniel EK and Martin Lorentzon in Sweden. It allows users to either stream music for free or stream music through a paid premium account. Premium account users do not have any advertisements and can even listen to music while offline. These streaming services can be used on a personal computer (PC) as well as on a mobile phone/device. Spotify is currently the “best option available for an on-demand streaming music service” (Hayes, 2016) due to its free access and user friendliness. Furthermore, Spotify is available in more than 60 countries globally. Nevertheless, Spotify also experienced some setbacks for instance, in 2009 “Spotify announced that there had been a security flaw in the Spotify service” (Casinos, 2011), that caused private user information to be potentially exposed. Additionally, some artists are being critical towards Spotify because they perceive that the compensation they receive for their music is low. Consequently some artists like Taylor Swift do not want their music on Spotify.

Key Resources and Channels:

Spotify, at its heart, has a massive and growing data set. The data they collect is mostly user-centric. For example, they can use information about a user’s activity and personal background like the location, what artists they listen to and like based on the Facebook profiles of most Spotify users, at what time of the day and what their friends listen to,etc. These information are used to generate recommendations, an understanding of tastes and preferences of users and groups which enables the service to develop playlists based on a user’s preference. This is where Big Data comes into the picture. It creates a service that feels personal for millions of people and analyses each and every one of those people. At the same time, by having such a large collection of music, Spotify can provide interesting analytics of music. For example, if a user wants to listen to a specific kind of genre or type of music, Spotify uses interesting algorithms and machine learning techniques to find the right kind of music for that person. Just to give an idea, in total, Spotify has over 28 petabytes of storage spread over four global data centres and collects about 4 terabytes of user data every day.

4Vs of Spotify’s Big Data:

We will now look at what makes up Spotify’s big data analytics from the point of view of four characteristic features.

Veracity

Spotify strives to be entirely data driven, ie. deliver insights and make decisions entirely based off data. Spotify’s idea behind behind data driven is that it is important to rely on data as much as possible to make decisions without human biases. Veracity (trustworthiness) of data is necessary in this context. Since Spotify collects most of its data from user activity and information over the internet, the data they collect can generally be considered reliable.

Volume and Velocity

In total, Spotify has over 28 petabytes of storage spread over four global data centres and collects about 4 terabytes of user data every day. This is huge amounts of data and requires analytics using technologies like HADOOP, particularly HDFS, which they can use to perform faster and distributed computing. In fact, Spotify has the largest HADOOP cluster in Europe. Spotify also uses Kafka for collecting data and databases to analyse and visualize it.

Variety

Spotify uses a variety of structured data from users and groups of users, such as: number of songs played by a user in a day, number of users listening to a particular album at a time etc. They can then use this to generate new insights such as predictions of what a user would like to listen to at a certain time during the day, what new artists a user might like etc. In general, they deal with a large variety of data that is useful to their analytics.

Key Partners and Customer Segments

By far, Spotify’s largest business partner is Facebook. Facebook was an integral component of Spotify’s first entry into the US market in 2011. Today, Facebook remains to be Spotify’s best sources for advertisement and new user acquisition. The integration between Facebook and Spotify is very sensible and intelligent. Spotify users have the option to link their Spotify account with their Facebook account. In doing so, they now have the capability to share songs, albums, artists, or playlists they are listening to. They also can now see the music their Facebook friends are listening to. However, shared music is not only shown to other Spotify users, but also non-users that are on Facebook. For Spotify as a business, this partnership has the advantage of both creating a more compelling and interactive product, and also helps expose their product to the larger market of Facebook users. Facebook also gains considerably from this partnership although in a less obvious manner. By integrating Spotify into the social media platform, Facebook generates additional user activity that makes its product more compelling to use. It also importantly gains more user information that can be used to better specialize advertisements and therefore make advertising space more valuable.

sp-fb.png

Source: https://valuechaingeneration.com/2015/06/14/how-spotify-experiments-with-partnership-enabled-business-model-innovation/

More recently adopted partners of Spotify include Starbucks and Tinder. Starbucks’ partnership with Spotify works by offering free Spotify premium accounts to Starbucks retail locations. In return, Starbucks creates and promotes Starbucks sponsored playlists on Spotify. These playlists are advertised both in store locations and online. The partnership creates advertising and promotional space for both companies. Spotify advertises its service in the retail domain of Starbucks locations, and Starbucks gains the ability to promote its brand to Spotify users (and consequently Facebook users). The online dating service Tinder’s partnership with Spotify functions in much the same way as Spotify’s partnership with Facebook. Tinder users can display their music preferences and likes on their profile with Spotify links. Tinder obtains a more compelling service, and Spotify can better access new markets and users.

As one might suspect for an online and socially interactive music service, Spotify’s user demographics trend very young. Millennials are responsible for 72% of music streaming across mobile and desktop usage. This equates to approximately 1.3 billion songs per week. Despite this, Spotify continues to advertise to younger audiences rather than older ones. Spotify currently controls around 41% of the paid music streaming market but faces considerable competition from other streaming services like Apple Music, Pandora, and Tidal. As could be concluded from the company’s aforementioned partners, Spotify continues to advertise in younger markets as a means to draw users to its service instead of those of its competitors. This is thought to be a more beneficial strategy than to attempt to attract older demographics who will likely not use streaming services at all in most circumstances.

Value Propositions:

The value propositions of Spotify are that it enables customers to listen to music. There are two ways that make this possible. The first and preferred way that the customers can listen to music is, when they have a free account which is a big advantage. However the disadvantage that comes with it, is that the customers are forced to listen to advertisements without being able to skip them. These advertisements appear every thirty minutes. Another way of having access to the whole database is by making a premium account, where the customers buy the right to stream songs for a period of time without having to listen to any advertisements. Compared to other music streaming services, at Spotify customers pay roughly 5-10 Dollars monthly to listen to more than 12 million songs without being forced to listen to advertisements counter to free accounts. At other music streaming services the customers need to pay a much higher price just for one song out of a whole album. In conclusion the value for customers is that they exchange their owner rights to any of the songs in order to have access to all of them but they don’t have to make a choice between which song to buy or listen to because they can access all of them.

screen-shot-2016-11-28-at-22-14-00

 

Customer Relationship:

Spotify’s customer relationship is a big success because the relationship between Spotify and its customers enables customers to build a relationship among themselves by sharing their favourite songs and artists with their friends who then also become Spotify users and share their music further with their friends with the consequence that this whole process benefits Spotify. Spotify has 75 million active users and over 20 million subscribers in over 15 countries. In order to track what every customer and every song that is playing, Spotify uses Salesforce Customer Relationship Management to manage a growing file of tracks that is being delivered to millions of listeners in over 15 countries. Spotify’s goal is to make music accessible to everyone and connect with everybody by using social networks such as Facebook, Twitter and Tumblr in order for the customers to share their music and their favourite artists with all their social media contacts which enables Spotify to be used further by just music being constantly shared. This approach is also the backbone of how Spotify runs its business. According to Jonathan Forster, VP of Sales, EMEA “Connections are important to [the] business”. Spotify is constantly asking its customers for feedback caused by how fast they are moving so that it would be a disadvantage to wait a year for the feedback of customers. The company’s main focus is though to first ‘’connect all the people inside the company.” [Martin Lorentzon. Founder and Chairman]

title.png

Cost Structure and Revenue Streams

Spotify services two types of accounts: free advertisement based accounts and premium subscription based accounts. Free users experience advertisements between every few songs where premium users pay a $10/month subscription fee. As one might expect, the number of free accounts well exceeds the number of premium accounts. Following the end of 2015, Spotify had approximately 160 million users of which only 39 million were premium subscribers. However, despite the disparity between the number of free and premium users, 90% of Spotify’s revenue comes from its subscription based accounts. Spotify is working to change this however and make its free users more profitable. In 2015, the company spent 250 million USD on researching user behavior and advertisement specialization. Spotify hopes to make increase the value of the advertisement space it offers in order to increase revenue.

screen-shot-2016-11-28-at-21-43-23

Source: https://www.bloomberg.com/gadfly/articles/2016-08-30/spotify-can-t-live-on-that-10-a-month-subscription-fee

Although profits have been found in some specific markets, as a whole, Spotify has not made a profit since its founding in 2006. In addition to the company’s largely unprofitable free user base, this also due in large part to the considerable costs associated with music streaming. The greatest cost by far is that of music licensing. In 2015, 55% of Spotify’s sales (1.8 billion USD) went towards covering licensing costs. Currently, Spotify is negotiating with major record labels to reduce these costs before the company attempts to become public in 2017. It is thought that reducing licensing costs to fall below 50% of sales would significantly improve the value of Spotify’s IPO. Another large cost that Spotify must cover is music streaming itself. In 2009, Spotify’s 5 million users streamed an average of 420,000 gigabytes of music per day. In 2015, Spotify’s now has approximately 160 million users who can also stream at higher bit rates than in years past; they likely stream over 14 million gigabytes of music on a daily basis. This, in addition to an ever expanding music library that must be physically stored, contributes to Spotify’s costs of operation and helps explain how such a large company can remain unprofitable.

screen-shot-2016-11-28-at-21-46-53

Source: (https://www.bloomberg.com/gadfly/articles/2016-08-30/spotify-can-t-live-on-that-10-a-month-subscription-fee)

SWOT Analysis:

Strength

Spotify’s main strength is its accessibility, it can be easily accessed from various devices, “it lets you play just about any song instantly without having to purchase or download it”(Chandler,n.d.) and it is user friendly. Consequently Spotify has currently about “100 million active users” (McIntyre, 2016) and 2 million artists with more than 20 million songs which makes it very popular among users.

As spotify is paying royalties to the artists and right holders the music on spotify can be streamed legally for free. On the other hand, Spotify tries to generate revenues through advertisements. However, users who do not want advertisements can create a Premium account and pay 9.99 € per month.

Weakness

One major weakness of spotify is that regardless of it’s many users it is not profitable due to its free streaming service.  Spotify has about 30 million paying users meaning they cannot pay high royalties. Furthermore, the “Music library is still limited by artists” (Hausman, 2013).  Some artists do not want to publish their music on spotify.

Opportunities

The music streaming industry has a large number of potential users and there is still a lot of potential to further expand digital streaming services. Moreover Spotify’s services of streaming music for free makes users who might have been used to “illegal or unreliable options for their digital music needs” (Hausman, 2013) switch to using their services.

Threats

Spotify’s major threat is their business model. Since they are not profitable, they can only pay small royalties to artists so artists like Taylor Swift who can make more money by selling albums will either continue to leave or not join their streaming services. If more artist continue to leave Spotify the number of users will drastically reduce, as their music needs will not be met by spotify. Moreover, David Lorry filed a $200 million lawsuit against Spotify due to copyright infringements. Additionally Spotify has a large amount of strong competitors like YouTube, AppleMusic and Amazon.

This shows that although Spotify is facing tough competition and has some weaknesses in its business model it has still a lot of potential to grow in the future. Besides the digital music streaming industry is a still expanding.

Conclusion:

In conclusion, we chose Spotify for our Big Data Project because we are all Spotify users, enjoying their services on a daily basis. Our goal with this project was to get a deeper insight on how Spotify operates, which challenges they are facing and how they use our information for their service. Furthermore we wanted to introduce the class to the Big Data of Spotify on how it works, what role it plays, and what its strengths and weaknesses are, as well as to address what kind of opportunities and threats it faces.

References:

Casinos, F. (2011). It All Makes Sense. It All Makes Sense. Retrieved from https://allmakessense.wordpress.com/.

Chandler, N. (n.d.). How Spotify Works. How Stuff Works Tech. Retrieved from http://electronics.howstuffworks.com/gadgets/audio-music/spotify4.htm

Hausman, A. (2013). The Spotify Business Model: No Guts No Glory. Capitalist Creations Entrepreneurship, Passions & Politics. Retrieved from http://capitalistcreations.com/the-spotify-business-model-no-guts-no-glory/

Hayes, T. (2016). The best music streaming service Spotify. The Sweet SetUp. Retrieved from http://thesweetsetup.com/apps/best-music-streaming-service/

McIntyre, H. (2016). With 30 Million Users, Spotify Is Gaining Subscribers Faster And Faster. Forbes. Retrieved from http://www.forbes.com/sites/hughmcintyre/2016/03/21/with-30-million-users-spotify-is-gaining-subscribers-faster-and-faster/#3fe709c6402e

Resnikoff, P. (2016). How Spotify Crushed a $200 Million Songwriter Lawsuit…. Digital Music News. Retrieved from http://www.digitalmusicnews.com/2016/07/12/spotify-crushing-david-lowery-lawsuit/

Spotify Solvay Case.(n.d.)The Business Model. Retrieved from https://spotifycase.wordpress.com/5-the-business-model/

Salesforce tops Spotify’s Playlist. Retrieved from http://www.salesforce.com/customers/stories/spotify.jsp

Doorneweert, Bart. How Spotify Experiments with Partnership-Enabled Business Model Innovation.. June 14, 2015. Retrieved from https://valuechaingeneration.com/2015/06/14/how-spotify-experiments-with-partnership-enabled-business-model-innovation/

Starbucks Announces Spotify Partnership. Billboard Staff. 19 January 2016. Retrieved from http://www.billboard.com/articles/business/6844194/starbucks-spotify-partnership-app-integration

McIntyre, Hugh. Millennials Are Leading The Spotify Revolution. Retrieved from http://www.forbes.com/sites/hughmcintyre/2016/04/20/millennials-are-leading-the-spotify-revolution/#75b56bb12f20

How much does Spotify cost to run? The Guardian. Retrieved from https://www.theguardian.com/technology/blog/2009/oct/08/spotify-internet

Spotify Can’t Live On 10 Dollars a Month. Bloomberg. Retrieved from https://www.bloomberg.com/gadfly/articles/2016-08-30/spotify-can-t-live-on-that-10-a-month-subscription-fee

Ingham,Tim. Spotify revenues topped 2 billion Dollar last year. Retrieved from http://www.musicbusinessworldwide.com/spotify-revenues-topped-2bn-last-year-as-losses-hit-194m/

Danielle, Jabin. Data at Spotify. Retrieved from http://www.slideshare.net/dj4b1n/data-at-spotify

Palmer, Jason. Analytics at Spotify. Spotify Labs. Retrieved from https://labs.spotify.com/2013/05/13/analytics-at-spotify/

Advertisements

6 thoughts on “Big Data: Spotify

  1. Very interesting insight. I was very surprised to read that Spotify has not been making money this whole time.

    One feature I really like about Spotift is how they come up with recommended music and more often than not, I actually happen to enjoy the songs that they recommended. In my opinion, this is an amazing way for me to discover new music.

    Like

  2. Another weakness Spotify has is its location-specific usage. The app does not allow you to use it for a period of time longer than 10 days, if you are outside of the country where you created your account or purchased your subscription. This is clearly incompatible with the generation of young people it seeks to target, as the individuals from this customer group have born and raised in the era of globalization, and thus are used to and demand accessibility to their apps, no matter where they are in our increasingly connected world.

    Especially because this innovation was developed in Sweden, one might think that its creators would be more aware of such issue, given the open-border policy of the European Union, and the easiness of young people to travel across borders.

    I have personally experienced discontent with this feature of Spotify, because when I go home during Winter Break, I am unable to access my personal account after 10 days of arriving.

    Like

    1. I completely agree with you! In addition also Spotify’s music rights depend on your location meaning that certain songs are only available from certain locations. This can create further frustration when some songs you really like are not available anymore and constitutes another weakness

      Like

  3. Nice article =)
    Like mchairani I´m surprised how well an algorithm can predict my music taste. Music taste seems to be something personal, unstable and emotional. So for me it´s incredible that spotify and Bandsintown can suggest so good.

    Like

  4. Recommendation algorhythms are often tricky to narrow down, with Netflix’ movie recommendations being notoriously bad. I think music is a fortunate case because it’s easier to follow genres, charts etc. but it’s still interesting to see how Spotify have leveraged big data to improve this area.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s