Today I want to discuss a strange paradox in modern music. Never before has it been so easy to release a song. Never before has such a huge portion of the recorded music catalogue been available to listeners instantly. And yet, for an individual artist, being discovered seems harder than ever. Why?
Let us try to look at the problem mathematically.
- The Algorithmic Discovery Problem
Suppose that approximately 100 million new songs are uploaded to Spotify every year. This number is probably inflated. A significant portion of these uploads are AI-generated songs, experiments, abandoned projects, or releases from artists who are not actively trying to build an audience. However, let us remove this noise completely and make the calculation more meaningful. Assume that an average serious artist releases about 10 songs per year. This gives us approximately 10 million artists releasing music every year. However, an artist does not need all 10 songs to succeed. Usually, there is one song that becomes the entry point — the song that introduces the artist to new listeners. Therefore, for the purpose of discoverability, the competition is approximately 10 million songs.
Now let us imagine an ideal world where Spotify recommendations are completely fair. Every one of these 10 million songs has exactly the same chance of being recommended. The probability of one particular song being selected would be:
Probability of discovery = 1 divided by 10 million.
Assume that a listener receives and listens to 50 recommended songs per day. The probability of one particular song reaching one listener becomes:
Probability of reaching one listener = 50 divided by 10 million.
Now multiply this by approximately 200 million Premium subscribers:
200 million listeners × 50 recommendations per day = 10 billion recommendation opportunities per day.
Under this completely equal system, a single song would receive roughly:
10 billion × 1 divided by 10 million = about 1,000 streams per day.
Of course, this calculation is somewhat favorable to the listener-discovery model. It assumes that old releases do not compete, that every new song has an equal chance, and that algorithms are not biased toward established artists, popular genres, or songs that already show engagement. Under these simplified assumptions, the mathematics suggests that an unknown song should receive meaningful exposure.
But this is not what happens. The overwhelming majority of independent releases receive nowhere near this level of attention. Therefore, the reality of music discovery tells us something important:
The algorithmic world is not a fair lottery where millions of songs compete with equal chances. Exposure is concentrated among a relatively small number of tracks, while the vast majority of songs remain almost invisible.
The calculation does not tell us exactly why this happens. It may be caused by recommendation strategies, listener behavior, previous popularity, marketing, playlists, or a combination of all these factors.
But it does show one thing clearly:
Simply uploading a great song and waiting for the algorithm to discover it is not a realistic strategy.
2. The Human Curator Problem
Perhaps algorithms are not the answer. Maybe human curators are the solution. After all, a person who actively listens to music should be able to discover something special. Let us look at the numbers.
Imagine a highly active independent curator who listens to 100 new songs every day. That is approximately 30,000 songs per year. Suppose this curator is serious and reasonably selective, adding 10% of the songs they hear.
This means that approximately 3,000 songs receive playlist placements annually.
Now we need to remember one important assumption. The competition begins only after every recording artist has already done the same thing: chosen their strongest song and tried to put it in front of curators. Therefore, the race is still between roughly 10 million songs that are actively trying to be discovered. The probability of a random song being heard by this curator is:
Probability of being heard = 30,000 divided by 10 million.
This is about 0.3%.
The probability of being both heard and selected is:
0.3% × 10% = about 0.03%.
This calculation already ignores musical preferences. A rock curator will not select a jazz song, and a metal curator will not select an acoustic ballad. Real-world matching makes the odds even more specific.
Now assume that a successful playlist placement generates about 5 streams per day. The expected result from one random submission is:
0.03% probability of placement × 5 streams per day = 0.0015 expected streams per day.
In practical terms, roughly 700 – 1,000 random submissions are needed to generate an expected additional stream per day. Even if every submission costs only $1, this becomes an expensive game. You may spend thousands of dollars chasing a very small amount of exposure. Again, the problem is not necessarily that curators are dishonest or that playlists have no value. The problem is that attention is limited. A curator can only listen to a tiny fraction of the music that exists.
3. Escaping the Random Game
Both previous calculations share the same hidden assumption: the artist is anonymous. The song enters a giant pool of competitors and waits for a random chance to be noticed.
This is where many promotion strategies go wrong. Sending more random submissions does not solve the problem. It only increases the amount of noise. The solution is to change the probability model.
Instead of sending music blindly to hundreds of curators, an artist can first understand where the music belongs.
Listen to playlists.
Study curators.
Learn their taste.
Find genuine connections.
Submit music only when there is a real possibility that someone will appreciate it.
The goal is not to increase the probability of acceptance. The goal is to increase the probability of being considered. Once someone actually listens, the music can succeed or fail based on quality, originality, and compatibility. But before that moment, the artist is not competing in a musical competition. The artist is competing for attention — and attention is the rarest resource in modern music. The paradox is simple:
Anyone can release music.
Anyone can reach the world.
But because everyone can do it, being genuinely heard has become harder than ever.
Music Picks
This time, I want to propose an experiment.
Below you will find 5 remarkable songs. The artists I chose currently have very modest monthly listener numbers. They did not ask me for this, and I have no incentive to select these particular songs. I simply believe they deserve to be heard.
Pick one song you like and spread the word about it. Let’s see if we can help these artists grow their audiences through the simplest and oldest music-discovery tool we have: one person sharing a song with another.
It will take only a minute of your time to tell a friend about an awesome song you discovered.
Let’s do it!
Here they are:
Mind Full of Bees by Jordan Branch:
https://open.spotify.com/track/4e0jzLYyyH4eJl5eK5YFNi
The Real One by Iam4given:
https://open.spotify.com/track/1DDo2GxMJSbPeNHhDceFde
L.A. Times by Westcoast P:
https://open.spotify.com/track/0BbfMnkPN4wWym7pvkV9DM
Got Me by Zasha:
https://open.spotify.com/track/0MQQHfy8IcZVgsw8Ratd5u
Euterpe by Flumana:
https://open.spotify.com/track/7ysiGppZOZD3Muwk6AkubQ
Final Thoughts:
There must have always been a bottleneck holding musicians back in some way. At first, instruments and musical education were scarce. Later, sound reproduction itself required massive resources. Then radio and television created bottlenecks of their own. And now, when modern distribution has essentially solved the problem of reaching listeners, it has simultaneously created a new one: the bottleneck has moved to the discovery process itself.
In this article, I tried to explain once more, from a different angle, the idea that I have been advancing since I started my musical journey: discovery was, is, and will always be based on meaningful human interactions and direct connections with those who might appreciate your craft.
It is not random. The calculations above explain why. It is not a matter of chance. And it is not only your hard work that will eventually bring you recognition.
Break the cycle of randomness. Take the time to know the people you want to reach and understand why they might appreciate your music. Only then, in an environment built on mutual respect, can a genuine "group effort" help you break through the vacuum of anonymous obscurity.
I have received thousands of songs and have reviewed up to 200 songs in a single day. More and more often, I see artists relying on randomized submission algorithms instead of taking the time to check the playlists they are submitting to and understand the people behind them.
This randomness is not only counterproductive for the artists submitting music; it also consumes valuable time and blocks the submission queue for those whose music might actually be a meaningful fit.
I wish to see this pattern change. I hope you agree, and if your fellow artists ask you for advice, perhaps spread the word about these good practices.
Wrapping up this post, I’d like to leave you with one simple but interesting calculation.
Imagine that you know 20 people well enough to suggest music you genuinely like. Now imagine that each of these people has another 20 people they can reach in the same way. After only 7 steps in this human chain, the number of connections becomes enormous — potentially enough to reach the entire world.
Of course, this is not an exact science, but the idea is still very telling. It reminds us about the power of human connection, honest communication, and sharing musical tastes with people who might genuinely appreciate them.
Think about it! I’ll be waiting for you here next time.