So you’ve just released a new album and are wondering why it isn’t getting as many streams on Spotify as your last one? It doesn’t always mean that people don’t like your music – more than likely it’s because they don’t even know about it.
The Spotify algorithm is complex, but once you understand the basics of how it works (and what to do about them), you can start to use that knowledge to build your audience of listeners and increase your streams!
After doing dozens of hours of research and even discussing with a former employee at Spotify, I’m going to share with you a handful of things that could affect whether or not someone hears one song over another. Based on these factors, your music could get more streams from Spotify radio, algorithmic playlists, or suggestions to listeners.
The Spotify Algorithm Explained By Music Industry Expert, Zac Harding
A couple of years ago I found this posted comment about the Spotify algorithm and was blown away by the amount of helpful information. It was by Zac Harding, who is a marketing consultant and music industry veteran. Because he explained it in better words than myself, I’ll add the entire quote here:
“With Spotify, the goal is to create a strategy that uses their algorithms to get your music on Release Radar and Discover Weekly playlists. How is this done?
First, they categorize music by putting a SongID to each song and a UserID to each person. They cross reference the SongIDs with the UserIDs to create a matrix of what people like to suggest songs.
Using this matrix, they use an artificial intelligence called Natural Language Processing or NLP for short. NLP technology combs through authority sites and blogs to see what people are writing about certain songs. So, if someone writes about your music on the site and relates it to Linkin Park or something then the NLP Technology takes that back and applies those relations to your SongID helping it to suggest your music more. It also tracks how much your music is being written about so if it’s really relevant due to let’s say a release then NLP will tell Spotify, “Hey, suggest this more now in Release Radar playlists.”
Then NLP runs through what they deem as “good” playlists on Spotify to see if these playlists are picking up your song as well. If so, they boost suggest your music more.
The first two learning models, Collaborative Filtering and Natural Language Processing, need streaming or text data about a song in order to properly position the song in latent vector space. Without a vector, or a “position” in vector space, Spotify can’t understand whether a given user will like a given song.
Because of this data requirement, any newly released song on Spotify is unlikely to be recommended if the song has not yet been streamed or written about online.
Spotify’s development team refers to this problem as the “Cold Start” problem. They have been working on solving this issue for many years.
As a partial solution to this issue, Spotify began analyzing the audio data (“spectrogram”) of new tracks.
Spotify decides each song by key, tempo, rhythm, time signature, timbre, and over a dozen others. This is then added to the SongID. This helps the song if there is no stream data or articles written online. But if you have all of those it only increases it’s suggestions of that song.
So, how can you leverage this?
- Make sure your metadata is accurate and thorough.
- Make sure your website has your biography, track descriptions and accurate sound descriptions of tour music. Like listing the artists whose fans you think would like your music and using terms that those artists use to describe your music.
- Encourage as many fans as you can to follow your profile. I’ve had success using a re-save campaign with a referral platform like Gleam.io to get followers for Spotify accounts and social media accounts on top of them referring their friends to do the same and follow plus pre-save. You also gather emails this way which can be used later in lookalike Facebook marketing. Also, setting weekly releases that encourage your fans to follow along and inviting them through social media and email.
- Speaking of ads, having a monthly budget that uses these lookalike audiences then makes sense because if you can keep the pressure up on new followers and streams Spotify will start thinking of adding your music to Discover Weekly playlists. But you need that email list of your fans first!
- Use Hunter.io to gather emails for top bloggers and journalists from all the top sites then start a drip campaign sharing your music releases with them. The more action you get her the more the NLP Technology will recognize your music is blowing up.
- Whenever releasing songs to Spotify, always upload to your distributor at least two weeks before your actual Release Date. Then Submit The Song for playlisting at LEAST 7 days prior to your release date by:
-Navigating to artists.spotify.com
-Clicking “Music” > Upcoming > Submit This Song (next to the track)
-Filling out the submission form using accurate data”
Factors That Can Affect the Spotify Algorithm
Please keep in mind that just like the algorithms on all major tech platforms (Google, Facebook, Instagram), these algorithms get very complex and we can never be 100% sure what’s included or not excluded. These are our best guesses (along with data & research) of what it leverages to prioritize your songs against others.
Release date
The more recent the better! New music is always prioritized over old songs, so make sure to put your new releases on Spotify as soon after they’re released in order for them not be lost among older content and forgotten about by listeners who are looking through their playlists or searching “new” albums.
The number of times a song has been added to listeners’ playlists
Do what you can to get fans to not just stream your tracks, but to actually add them to their playlists. This will help you get more streams AND tell the algorithm that people truly like your songs enough that they want to listen more in the future.
The number of times a song has been liked, saved, or downloaded
Strategic artists look closely at their saved-to-listeners ratio to see how well people are interacting with their song. We don’t know for sure, but I’d guess the Spotify algorithm also leverages this data.
The number of times a song has been shared on social media.
It’s always a good idea to request your fans to share your track on their Instagram. Even if your fan group is small, just getting friends and family to do this will teach the algorithm that people like your music enough to share it.
How many followers you have on social media, and how often they interact with your posts (likes, comments & retweets)
This is also important for building an audience because it teaches Spotify that people are interested in what I’m posting!
Listener’s Location
A listener’s location, which can be determined by the language they’re listening to or their IP address (if it matches with one in an area where your music is popular).
What else is there?! Algorithms for big tech companies like Google, Facebook, Instagram, Spotify, etc are always changing and updating. By the time you read this there may be new factors at play as well. Let us know what you think in the comments!