Millions of people around the world now get their music from streaming services. Ten years after its debut, Spotify, the market leader out of Sweden, now counts more than 75 million users, of which around 30 million are paying subscribers.

Other music services like Apple Music, Napster and Deezer in the meantime have been able to convince millions of people to pay for their music each month. For a small fee, customers can access an enormous music library with more than 30 million songs available at the bigger services. But just how do you make your way through such an enormous offering?

Algorithms look for more

"The days spent thumbing through vinyls at the record store have passed; this cultural practice is dead," said Stephan Baumann of the German Research Institute for Artificial Intelligence. The IT expert is himself a DJ and recalls fondly the many hours he's spent at his favorite record shops. He's not mourning the lost days, but instead looking pragmatically at how remarkably access to music has changed.

The question of how users discover the music they love when stumbling through a mass of offerings is something that streaming services have responded to with algorithms. These algorithms analyze a user's taste based on the selections they've made so far and automatically present a new song that fits that taste.

"Algorithms can find the pearls that are uploaded somewhere in the world and liked by a few and that we as individuals would have otherwise overlooked," said Baumann.

The right mix is a question of emotion

For more than 15 years, Stephan Baumann, a leading expert in his field, has been researching this system of recommendations, which is one of the disciplines covered by artificial intelligence. Yet the weakness of this form of recommendations is its lack of ability to display or respond to feelings.

For many people, as Baumann known, music is an emotional experience. "As long as your life history isn't contained in your computer, the algorithm remains emotionless."

Finding the right mix is not so simple. There is also the problem of the "filter bubble," which is created as a result of these algorithms. The user only receives recommendations for music that he might like based on what he has already listened to. It doesn't take others' opinions into consideration. Over time, this may threaten the streaming services as the music selection narrows and lacks surprises.

Stephan Baumann looks at the situation dryly. "We've all made ourselves comfortable in our little filter bubbles. If I'm an active Facebook user, then I accept the fact that certain things won't appear in my timeline, which actually came from my friends."

It's the same case with music found via streaming services. To make up for these limitations, the user doesn't have to slug through alone for good music. In that respect at least, the algorithms can be helpful despite their weaknesses.

What's your experience with streaming services' recommendations? Share your thoughts with us on Twitter: @dw_culture.