Legal and Tech: How On-Demand Music Streaming Works

Everybody loves music streaming services; they provide free and frictionless access to any music whenever and wherever we want to listen to it. For the past few years, streaming services have seen steady growth and, according to Statista, streaming services are the most profitable digital music format in the US.

The tech giants are battling for position in the $25.3 billion per year industry along with hundreds of startups seeking their own niche. On the battlefield of music streaming domination are Amazon Prime Music, Apple Music, Google Play Music, Deezer, Pandora, SoundCloud, Tidal and many others.

In this article, we decided to examine how the streaming market works. This analysis will help answer the question of how you might enter this market with your own streaming solution.

Types of streaming services

There are two major types of streaming services: radio and on-demand. Radio stations are mostly associated with music discovery, with an app acting as a DJ who picks what songs to play and in what order. On-demand services are characterized by user-created playlists that can be shared with others. Let’s talk about these two types of streaming services in more detail.

Radio stations

The best representatives of radio apps are Pandora and iHeartRadio. These apps aren’t intended for listening to the one specific song that you want to hear right now. These services call themselves “radios” because they play songs you might like based on your personal preferences.

For example, if you like alternative rock, you can type in Papa Roach and get songs from bands like Shinedown and Breaking Benjamin. You can’t select a specific song with on-demand radio stations like Pandora, but you can create your own radio stations based on genres, songs, albums, artists, or bands.

Radio stations have licensing agreements that don’t allow them to play songs on demand. These licenses allow radio stations to keep their costs down compared to on-demand streaming services.  

On-demand streaming services

Examples of on-demand music streaming services include Spotify, Apple Music, and Tidal, all of which let users play songs of their choice instantly. Similar to radio station streaming services, on-demand streaming services also offer recommendation and discovery features based on what users like to listen to. And the other way round.

Given the growing market demand for on-demand music, Pandora just launched Pandora Premium, a $9.99 per month streaming service that competes directly with apps like Spotify and Apple Music. Pandora’s Premium service allows users to listen to Pandora radio while also creating their own playlists.

To develop its on-demand streaming service, Pandora needed to sign licensing agreements with major record labels as well as independent distributors. As a result, about 70 percent of revenue from Pandora's subscription service would go toward royalty costs for songwriters and artists.

Speaking of licensing agreements, I bet you would love to know more about how they work.

how music streaming apps work

[The Pandora app. Image source: The Verge]

Licensing agreements

There are two types of properties that require contracts or licensing agreements that you need to sign to make a music streaming app: sound recordings and compositions (songs). Sound recordings and compositions come with rights that can be owned by multiple rights holders.

For example, if an artist signs a deal with a label and records his or her song in the label’s studio, then this sound recording belongs (at least partially) to the label.

The song itself (the composition) belongs to the songwriter, although a songwriter can have a publisher manage those rights. Managing rights to compositions means issuing licenses for the use of songs, collecting royalties, dealing with accounting, and managing other administrative issues. Very often, songwriters sell their songs to a publisher.

Every internet radio station and on-demand streaming service needs to pay copyright holders for the right to reproduce or make copies of sound recordings and compositions. These payments are called royalties.

However, internet radio stations and on-demand streaming services require different licensing agreements and must strike these agreements with different organizations.

Who must radio stations sign licensing agreements with?

Radio stations pay government-approved organizations that in turn pay rights holders: labels, music publishers, and songwriters. Radio royalty rates in the US are set by the Copyright Royalty Board (CRB). You can read more about the CRB here.

If you want to develop your own radio station, you need to get licenses from the following agencies:

1. SoundExchange for sound recordings. SoundExchange compensates labels and individual artists.

2. Performance Rights Organizations (PROs) including ASCAP, BMI, and SESAC for publishing licenses. These organizations pay royalties to publishers and songwriters.

Who must on-demand streaming services sign licensing agreements with?

To develop an on-demand streaming app you will need to license content from the following organizations:

1. Major labels such as Sony Music, Universal Music Group, and Warner Bros.

2. Independent aggregators such as the Merlin Network that represent indie artists.

3. Publishers such as Universal Music Publishing Group and Sony/ATV.

All these organizations require direct licensing deals, which typically take one of the following forms:

- per stream rate

- percentage of revenue

- upfront payment for future streams

- equity

For example, Spotify pays 55 percent of its revenue to labels (not including publishers). Apple Music pays 58 percent of its revenue to labels.

How to develop a music streaming app like Spotify

[The Spotify app. Image source: Vulcanpost]

As you can see, digital streaming services are a profitable distribution channel for record labels and publishers. The promise of income from streaming subscription services has prompted copyright owners to license the global expansion of services such as Deezer and Spotify, both of which stream audio content from Sony, EMI, Warner Music Group, and Universal Music Group record labels. However, labels have been rather reluctant to sign deals with emerging streaming services that don’t have a lot of users to profit from.

The rules on the streaming market aren’t very friendly for emerging digital businesses. To make things even more troublesome, there is no global standard for music licensing, and therefore music app development requires you to license country by country.

Dealing with boring legal matters isn’t the only challenge that a music streaming app developer needs to tackle.

The streaming market is shifting towards discovery and recommendations. To survive in this highly competitive digital industry, your app needs to provide a better listening experience than other apps.

Music recommendation and discovery

Services like Spotify and Apple Music are popular among users because they understand people’s preferences. This understanding is the result of analyzing considerable amounts of user data. Here are some methods that the largest apps use to develop algorithms that generate recommended playlists:

1. User onboarding

A new app doesn’t have large amounts of user data they can analyse. The first time a user logs in, you don’t know anything about them. To make a new user’s experience at least a little bit personalized, you can use Facebook’s API to get access to data about a user’s “likes” and other music-related information in their profile.

Another way to tweak your streaming to each user’s taste is to ask new users to answer a few questions about their favorite artists and genres, as Apple Music does in their onboarding interface.

2. Tracking user behavior

As your user starts listening to songs streamed from your app, you can let them like or dislike compositions they hear. When a user taps to like of dislike a song, you can save this preference to their profile along with the artist name, album name, genre, country, year of release, time of day, user’s age, and other information.

You can also analyze what tracks a user skips and which songs they listen to repeatedly. Such information can help your app to sort songs according to a user’s taste.

Gathering user preferences during onboarding and tracking user behavior over time can improve the listening experience. But these methods still don’t solve the larger problem of searching for similar music across the data in your database.

3. Algorithms

After you acquire a sufficient music database and user base, you can start thinking about developing a more advanced system for recommendation and discovery. To date there have been three approaches to curating recommendations in the streaming world:

  1. Audio analysis algorithms that automatically analyze all available tracks

  2. Human curation of music data  

  3. A combination of the two approaches mentioned above

Spotify, for example, which owns the Echo Nest, uses this company’s API. The Echo Nest uses data mining and machine learning techniques to gather information from the web about songs, albums, artists, and genres.

The technology that the Echo Nest provides can capture live playback behavior (artist plays, song plays, skips, bans, favorites, and more) and use this behavioral data to personalize playlists. The system generates playlists using similarity searches on cultural text data from the web, audio data from tracks, and data on user behavior.

The Echo Nest API also allows you to build social discovery features so users who share common tastes can discover music from each other.

Pandora’s recommendations, on the other hand, are based on the Music Genome Project, an automated musicological analysis that ignores genres, user connections, and ratings. The idea behind this project is to figure what you like (not what a market might like) by analyzing the musical structures present in the songs you like and then playing other songs that possess similar traits. This approach is definitely interesting, but it does distance users from the tastes of their friends, peers, and critics.

To develop their algorithm, Pandora employs trained experts who tag songs with dozens of pieces of metadata. These metadata include specific tonal qualities, instruments played, rhythmic nuances, and hundreds of other details. Pandora is conducting ongoing experimentation on its vast user base to figure out how best to deliver music to its listeners. You can read more about their intelligence algorithm here.

The streaming market isn’t the easiest to enter. To develop a music app for this market you need to have both a thoughtful business strategy and advanced technology.

Just for the record, the economic metrics for the ticket booking business look far more favorable than those for the streaming industry. According to Pandora, its Ticketfly platform that sells ticket to live events is forecasted to exceed $300 billion in revenue by the end of 2020.

4.2/ 5.0
Article rating
Remember those Facebook reactions? Well, we aren't Facebook but we love reactions too. They can give us valuable insights on how to improve what we're doing. Would you tell us how you feel about this article?
Music App Development Case Study

How we developed a service for iOS and Android where you can consolidate all of your music online

Read now

We use cookies to personalize our service and to improve your experience on the website and its subdomains. We also use this information for analytics.

More info