Introduction

Spotify Technology S.A. (/ˈspɒtɪfaɪ/) is a music streaming service developed by Swedish company Spotify Technology, which is headquartered in Stockholm, Sweden [Wikipedia]. Spotify operates under a freemium business model (basic services are free, while additional features are offered via paid subscriptions). Spotify generates revenues by selling premium streaming subscriptions to users and advertising placements to third parties [Wikipedia].

Revenue

Spotify platform includes 159 million Monthly active users and 71 million Premium Subscribers as of December 31, 2017 [prospectus, page 1]. For the years ended December 31, 2017, Spotify generated €4,090 million in revenue, respectively, representing a compound annual growth rate (“CAGR”) of 45%. Net loss in 2017 was €1,235 million. Cash flow was positive €109 million. For the years ended December 31 2017, Premium revenue comprised 90% of total revenue (€3.674 million). Ad-supported revenue was just €416 million [prospectus, page 70].
Average American listens to more than 32 hours of music each week [Nielsen report, prospectus page 2]. The average Spotify user listens 25 hours per month in Q4 2017 [FT]. Average revenue per Spotify user was €5.32 in 2017 claimed FT; we got a better number: revenue per user = €4.090 million : 159 million users = €25.72 p.a. or €2.14 per month. This is a very good value comparing to traditional radio. Assume that CPM of 30 sec. spot in the terrestrial radio is €2 and there are 10 spots in average hour (high-ranking station) so that the revenue per hour and listener is €2*10/1000=€0.02. Monthly revenue assuming listening 25 hours per month will be 0.5€ (€0.02*25hours). Spotify seems to be 4-times better due to the premium subscribers and more valuable advertising (on the first look).
In the United States alone, traditional terrestrial radio is a $14 billion market, according to BIA/Kelsey. The total global radio advertising market is approximately $28 billion in revenue, according to Magna Global [Source: prospectus, page 2]. There are 3.6 billion internet users globally [International Telecom Union ICT Facts and Figures 2017 Report]. 80%-90% of adults listen to radio daily in most countries [Statistics to radio consume]; same people are also Internet users. Therefore (if assumed that Internet users are music or radio listeners): Spotify share is just 159 million : 3.6 billion = 4.4%; if the total global radio advertising market is €32 billion (EUR/USD 1:1.15) then the Spotify revenue share is: €32 billion : €4.090 million = ~12%. This is also a good value that can be explained by premium users, more valuable advertising forms and Spotify presence on the mature markets with greater CPMs (at the first glance). Digital advertising spending worldwide was $229.25 billion [Statista] and typical user surfs 2-4 hours daily in Internet [ARD/ZDF onlinestudie]. If the Spotify share 4.4% of global Internet users than the advertising revenue from Internet users should be 4.4% from $229.5 billion budget = $10.098 billion. If we assume that the typical Spotify user surfs in Internet and listen to music from Spotify in a half of time (2-4 hours per day Internet usage but only 1 hour to Spotify listening) than we get something like: $10.098/2=$5 billion that is pretty close to the Spotify revenue.

Costs

Revenue is only one side of the medal. The other side is costs.
Spotify database consists of 35 million tracks [prospectus, page 3]. The vast majority of the content is licensed to Spotify on a non-exclusive basis. Spotify believes that personalization, not exclusivity, is key to continued success. In addition, because Spotify provides artists with access to millions of Users around the world, many artists independently choose to release content on Spotify, even if they initially provided this content exclusively on another Service [prospecuts, page 114]. Therefore, in December 2013, the company launched a new website, “Spotify for Artists”, that explained its business model and revenue data providing various benefits for creators, e.g. Monetization, Discovery, Distribution, Promotion, Analytics, Tools for Creation. The company pays 70% of its total revenue to rights holders [Wikipedia]. Just to verify this number: Spotify pays about $0.006 to $0.0084 per stream to the holder of music rights [CNBC report]. Assume that the typical user listens to 18 songs per hour; 25 hours a month and 12 months a year. In this case, the user listens: 18*25*12=5400 songs p.a. If Spotify pays $0.006 to $0.0084 per play to the holder than the payment per user must be: 5.400 songs p.a. * €0.007 = €37.8. We calculated previously that the revenue per user is €25.72. It seems that the license costs are greater than the revenue per user. Information is Beautiful provides an info graphic with slightly different numbers. According to their research: Spotify pays about 0.0038$ per play. 5.400*0.0038=$20,52 p.a. or €17,65. €17,65/€25.72=68.62% that is pretty close to the number mentioned by Wikipedia.

Competition

Spotify competes for the time and attention of users across different forms of media, including traditional broadcast, satellite, and internet radio (iHeartRadio, LastFM, Pandora, and SiriusXM), other providers of on-demand music streaming services (Amazon Prime, Apple Music, Deezer, Google Play Music, Joox, Pandora, and SoundCloud), and other providers of in-home and mobile entertainment such as cable television, video streaming services, and social media and networking websites [Wikipedia, prospecuts, page 28].

Current management strategy

If we look again in the prospectus than the management strategy is pretty clear:

  • Increase a number of premium users: Premium users are most valuable customers with doubling revenues compared to advertising users
  • Increase the customer/user base in the existing and new markets
  • Increase the average listening time providing more valuable service

Costs reduction

In order to reduce the royalty fees, the company is looking for new features and content sources:

  • Lyrics and behind the lyrics. Spotify provides textual and graphic information about the playing tracks
  • Podcasts. Spotify offers a growing number of podcasts, which have gained significant traction with users, creators, brands, and advertisers. Spotify platform enables seamless dissemination of podcasts covering a wide range of genres and topics, including musical content, sports, business and finance, travel, and cooking, among many others [prospectus, page 110].
  • Video. Spotify also have invested in expanding the content offerings to include video. Embedding video clips into activity-based playlists helps provide a visual layer to content offerings, allowing artists to promote their music further through graphics or photos on social media platforms while the song is playing. Integrating video into our most popular playlists, such as RapCaviar, Rock This, and Viva Latino, enables us to convert our listeners into viewers of content. Video content includes interviews, freestyles, behind-the-scenes footage, and full-length music videos [prospectus, page 110].

Stations by Spotify (transition from on-demand to streaming)

In January 31, 2018, Spotify launched a new Pandora-styled standalone app called Stations by Spotify. It features 62 music channels, each devoted to a particular genre. The aim is to get users to listen to the music they want straightaway without information overload or spending time building their own playlists [Wikipedia]. The second goal in this strategy is to reduce the license costs because music streaming costs are much lower than music on-demand costs, e.g. the typical FM music radio pays about 2% of revenues for music licensing [https://www.broadcastlawblog.com].

Spotify comparing to terrestrial radio

„Our mission is to unlock the potential of human creativity by giving a million creative artists the opportunity to live off their art and billions of fans the opportunity to enjoy and be inspired by these creators [Prospecuts, page 1]“. “Spotify is differentiated from other services because we provide Users with a more personalized experience, driven by powerful music search and discovery engines” [prospectus, page 3]. Spotify uses the user data to create a more personalized and engaging experience for each incremental visit. From the moment Users open the Spotify application, he/she gets a personalized homepage with content that reflects understanding of their music tastes, listening habits, musical moods, and daily activities [prospecuts, page 4]. Spotify believes that this personalized experience is a key competitive advantage as Users are more likely to engage with a platform that reflects the real-time moods and activities and captures a unique understanding of moments in their lives [Prospecuts, page 4] .
It seems that the management concentrates only on the music and artist side. Recently, Spotify includes Podcasts but podcasts are not a major revenue source and they are not integrated well. From the other side there is a classical terrestrial radio that will be seemed as a competitor from Spotify.
People listen to FM-Radio because:

  • Music: One of the big reasons people stay connected to their favourite FM-station is to hear music they know and love. [Radio.co]. They don’t need to search in the catalogues, playlists or other sources. All songs are chosen by the music department and tested by listener groups. Just to mention: typical radio station has less than 1.000 songs in program. Major cities have about 50 radio stations. Therefore, it is less than 50.000 songs that are playing on the active rotation. Music on-demand requires that the provider should include 100% of all possible titles but reality is: there are about 50.000-100.000 songs in a high rotation. The rest is a long music tail with few plays per month.
  • Radio is Part of Your Life; Over half of listeners say they tune into radio for local news, local weather, local sports. [Radio.co]. Spotify provides only music or long podcasts. It is great in some situation, e.g. party with dance music or high-concentrated work without interference factors but there are a lot of situations when Spotify can provide more background information injecting it into the playing music; it can be local news, short shows, user-generated content or some notification via TTS etc.
  • Radio gives a companion across their days [Quora]: Many people listen to stations with clever, conversational presenters and talkback because it gives them a feeling to not be alone across their days.
  • World news, breaking news [Quora]: it is very easy to start; just turn it on and have it in the background. Radio is there during emergencies and it will stay around for a long time.

Commercial FM-Radio is more attractive from a finance perspective; licensing costs are just few percent from revenue; there are only 12 songs in an average hour compared to 18 songs by Spotify. Assume: that the license costs per play remains same (what is not the case by streaming) but Spotify will be able to reduce the number of songs to 12 due to additional content in an average hour. In this case the music licensing cost reduction will be more than 30%. This is fast a billion EUR p.a. in the current numbers.

How to make Spotify profitable?

Make radio stations and their content first-class citizens

Radio stations are seemed to be competitors to Spotify or they are seemed to be competitors. As a result, most radio stations create additional streams parallel to the main channel providing alternative music or podcasts that were not included into the main program. It costs time, money and additional resources in order to establish such additional streams. Most dedicated Internet streams are just music with advertisement. It is possible and we successfully realised such projects to synchronise the main program with online streams providing a full program including the live hosts in all internet streams. It is possible and doable but as mentioned previously: it is not a main business of radio stations and most stations do not have a technical knowledges to implement it. Assume, Spotify could provide a standardised platform for radio station-makers with easy integration of the main program to alternative music. As a result: you can hear for example an Antenna Bayern in 30 music variants from Pop to Rock but with the same news, hosts, etc. powered by Spotify. In this case Spotify will achieve several advantages:

  • Additional listeners: people that listen to radio will do it on the Spotify platform
  • Reduce license costs: most radio stations already have a music license agreement and costs for streaming are lower than for on-demand services
  • Additional advertisement: Radio stations make comprehensive studies regarding the listeners and regions; all this information can be used for the advertising companies
  • Business model – revenue sharing: radio station makers will get let’s say 60% of the advertising

The cooperation will be attractive for radio stations as well because they get a technical platform with a musical content including background information etc.

Create a market place for non-music content

There are a lot of creative minds that can produce an original content and would like to share it with other people. Youtube is a good example for supporting custom production providing Youtube-channels with billions of views [Socialblade]. Typical FM-radio stations have a standardised pool of program elements, e.g. Jingles, News, Shows etc. Typical overlay to song is just 30 sec. long and includes only text that is relevant to the time of the day, current content or local or world news. Typical show is just 1.5 min long and includes all necessary elements from begin to end. Assume that Spotify establishes a platform for people that can distribute their content in the standardised way that will be injected into Spotify audio streams based on the preferences of the listeners.

The program can be mixed on the fly including the elements from news, shows, weather etc. The Spotify listener will get a personalised stream including own interest and preferences. It should be possible to make a market place in such way that the program elements can be combined automatically or on-demand, e.g. the listener is in Paris and would like to hear the French news but also like a comedy show from England; and all these elements have to be mixed with songs like in FM-radio. Additional content should also include video and text information. And it should be possible that the content creators will get a share of the advertisement if their content was played. It should be mentioned that the number of daily content producers will not be great: e.g. the content maker will need hundred thousands or millions of plays in order to earn a small income but it is imaginable that radio stations, studios and content creators can provide their content for the extra monetisation. Such content creators can be compared with artists but with standardised content. May be it makes sense to include a non-professional content makers as well, like a audio twitter with a prescribed forms of audio blocks so that listeners can subscribe to such channels a get a notification (or play) if new content is provided.

From finance perspective it can be very attractive as well: it will definitely increase the listening hours, attract new listeners and decrease music licensing costs because the new content can be licensed on the Spotify conditions. Moreover, this community will be locked to the Spotify service. It is not only 30 million songs with some algorithms for the play list generation. It will be a market place with a lot of people (commercial and non-commercial) that would produce a Spotify content every day for their listeners.

Provide a personalisation

Since all our Users have to log in to access Spotify Service, this enables Spotify to track behaviours such as playing songs, sharing, selecting recommended music, skipping, following, and active participation through the upvote and downvote buttons. More than 150 billion of these kinds of events are logged daily on the Spotify Service. Furthermore, Spotify devotes substantial resources to analysing and handling this data in order to obtain useful insights with machine learning and artificial intelligence. Spotify have the ability to personalize and curate the content by measuring an individual User’s preferences against more than 40 different parameters. Current algorithms are designed to anticipate a User’s preferences using factors such as demographics and past listening behaviour. Already, Spotify can combine situational context, such as time of day and location, to make better recommendations for appropriate content to an individual User based on his or her current activity [prospectus, page 115]. All this information is currently used for the better play list generation or music recommendation services. Assume, Spotify introduced a market place with original non-music content so that users will listen Spotify several hours per day. In such case, Spotify can include some AI TTS additions into the program that will be integrated into the music like in a traditional FM-Radio, e.g. instead of host voice-over between songs; the user gets a generated audio like: “You have an appointment at 10 a.m. in the city with Jack; Please don’t forget to reserve a table“ or “Paula posted new images on Facebook, take a look” or “You are driving E65 and there is a traffic jam in 2 km”.

It sounds impossible to realize and such service can be compared with Cortana, Alexa or Alice but it shouldn’t be the case for Spotify: Spotify needs only to expose an API for third-party developers that can inject their services into the main sound stream providing different useful data.

We did several tests with the traffic announcements that you typically hear in FM-radio (traffic or road situation after news block); in most cases this information is not relevant for listeners but it is well accepted and sounds great; we added this announcement to the Internet radio stream based on the device location (it was just a sound block that is relevant to geographical coordinates and time); it works and listeners have feeling that they have a partner that can help them in their current situation; please note the listener is listeneing to his favorite music radio station with a decent percent of this favourite songs and get the personlised recommendation on top regarding to his current situation.

From finance point: it will be a huge market with textual context-relevant advertisement. The advertisement spots will not include classical Audio/Video spots but also context-relevant text2speech announcements on the auction basis, e.g. if you are looking for restaurant; your map shows you 5 restaurants that are relevant to your wishes and location; after next song you get an announcement in the next song intro: “we have a special offer today in Italian restaurant “Pandora”; all pizzas cost only €4” and the vocal of the next songs begins. We make tests: if such announcement under 20-30 sec. it is well accepted by listeners and they not feel be disturbed like with classical spot and if it mixed as mentioned above, it sounds really great.

The advertisers can provide a real-time configuration of their budgets, e.g. switch off the spots if a restaurant is full or make (it is already possible by big agencies (RTB) but not for small advertisers). It is important to note: personalisation and personal advertisement must be developed together in other case the personal advertisement will be switched off due data protection regulations.

Summary

These are just few ideas that can be implemented by Spotify.

P.S. If the cross-fading options is on, the mix with the advertisement block begins on the end of the song so that a part of the advertising on Spotify is not hearable or people will not aware of the advertising spot; it was always no-go by the traditional radio. This is not a feature, it’s a bug.