The disconnect between music recommendations and music lovers. Part 2: API’s and Apps

Full disclosure: I am the founder of a music recommendation service, i will not bring that up until later. This piece is mostly focused on the general state of digital music recommendations.

In the previous post I discussed some of the issues of the major music providers recommending music. Both the issues with them trying to recommend 30 million different tracks to 100 million different users. And the shortcomings of the tools they use.

Wingclipped from the start

Being unable to cater to all users at all times some of the providers are offering developers the possibilities of interfacing their data to create different products which caters to different types of people. sometimes in creative forms (although for some reason never games).

This is important for the echo system of music since this opens up the possibilities of other types of data. Essentially making the providers platforms and not just products. Spotify appears to take this the most seriously and have a flexible API and SDKs which they actively maintains. Apple waited a year before realising that letting creative entrepreneurs access their data was probably a good idea. And even then they release nothing more than a remote control without any real control of playlists and similar tools. Soundcloud also offers an API but has a threshold of 15k requests per day making it tough to use it for other things than hobby projects. Deezer also have an API. Google play have unofficial APIs with no playback, Tidal has none.

YouTube, which still is the biggest on demand music provider does not permit 3rd party apps to play in the background and extracting audio is also against their TOS. This is however (of course) available in their own music app.

Terms and conditions do apply

The streaming providers all provide extensive terms and services which developers need to agree to in order to access to their content. Some of these are intentionally vague and breaking them may result in termination of access to the platform.

Examples of restrictions (some previously mentioned)

  • No background audio of videos (youtube)
  • No integration with 3rd party data (which I assume must apply to more or less all of the Spotify apps)
  • No integration with quiz or other type of games (which is not clear if it includes gamification)
  • No advertising or paid features
  • No caching of data (to ensure fast loading etc)

All of the provides have a long and large document with these restrictions. if you combine services which many apps do you run in to chances of breaking one agreement in order to comply with the other.


So armed yet fairly wingclipped, music loving entrepreneurs and developers have created hundreds if not thousands of music apps.

I will try to cover some of the issues some of these apps have.

Base content on your existing friends

This is a common approach to tackle the so called “empty room problem”. Meaning that you open the app and then have no idea what to do, having some friends around who have done something in the app helps, just follow them and you’re off. In music however this issue isn't as easily solvable.

Getting music recommendations from your friends is only good if you have friends with not only great taste in music, but also the same taste as you! If you don’t, then this is will only clutter, recommend things you already know and not solve the problem.

Sharing music isn't content creation

In Music the majority of visitors are consumers not creators. Unlike apps like Instagram where you share the photos you take yourself. Music to a large extent does not fit that format. Soundcloud and band camp technically does.

If anything music apps are more pinterest than they are instagram. sharing things you find and love rather than sharing what you create. “The Instagram of music” makes no sense. Otherwise you would only share music which has pretty art work.

Doesn't solve an actual discovery problem.

A lot of apps are focused on popular music, music which greatly overlaps with the Playlists and curated lists of content which are peddled by the major record labels. If this is the case the product will just like all other social apps rely on user engagement to create value. If it doesn’t have either it has no real justification for existing other than novelty.

Playback restrictions

Some apps base their content on one platform, some on many. These follow restrictions, like Youtubes no audio only and no playback when the app is closed. Or soundclouds hard limit on usage per day.

Even if you make it work with these restrictions the global user base are scattered between providers. Backing the wrong horses will lock out a major chunk of potential users. Making things even harder provider usagw also greatly differs between music styles depending on availability for said providers.


In defence of independent music apps

Diversity is important, especially for music as its increasingly being focused to a few major players, and probably even fewer in the near future. Yet the tools available are restricted both in features, access and against making profits in most ways.

I’ve often heard and discussed that the perfect music app hasn’t been created yet. this is most likely because the only ones with the complete access to make something great have to spend all their time to cater to the music business unhappiness of compensation.

note: There is also a somewhat growing trend of artists releasing their own proprietary apps (which is also against the streaming providers TOS if you are using them). I for the most part consider this a marketing tool and not something that will facilitate music discover or consumption.

In the last part of this series I will discuss how and why we tackled all of this in Refine music.

Like what you read? Give Johan Mickelin a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.