What to release next, and how often

Two decisions get made badly more often than any others, and both get made in the same way: whichever song the artist is currently most in love with, released whenever it happens to be finished.

Both deserve better than a mood.

Let twenty people pick the single {#twenty-people}

You cannot hear your own songs. You have heard the demo four hundred times, you know which bar was hard, and you have a private relationship with the one nobody else responds to. This is normal and it makes you the worst possible judge of which one to spend money on.

So do not judge. Test.

Get the candidates in front of about twenty people who are not your friends and not your family. A local group, a Discord, a few strangers at a rehearsal space, whatever you can reach in a week. Then watch what they do instead of asking what they think.

Behaviour that counts: they play it twice. They ask what it is called. They send it to somebody. They bring it up again a week later without being prompted.

Behaviour that does not count: “that’s really good.” People say that to be kind and it correlates with nothing.

The number that tells you the same thing {#the-number}

If a song is already out, there is a cleaner signal available, and almost nobody looks at it.

Take streams and divide by listeners. One person who plays a track eleven times and one person who plays it once both show up as a listener. Only the ratio tells them apart.

A track with fewer listeners and a higher ratio is a better track to push than a track with more listeners and a ratio near one. The first one is being kept. The second one was heard once and forgotten, and paying to show it to more people just buys more people forgetting it.

This is the single most useful number on your dashboard and it is not on the front page of any of them.

Four to eight weeks, decided in advance {#cadence}

Now the second decision, which is pace.

The long gap is the classic independent mistake. A track comes out, it does moderately, the artist goes quiet for seven months to make the next thing perfect, and by the time it lands every platform has forgotten them and so has everybody else.

A rhythm of one release every four to eight weeks holds a middle line. Close enough together that you stay present on surfaces that reward recency, far enough apart that each track gets a real runway rather than being buried by the next one.

What makes it work is that the dates are set before the songs are finished. A calendar you wrote in advance is a commitment. A calendar you fill in as things get finished is just a record of what happened.

And the honest part: the first several releases are not going to do much. That is not a sign the plan is wrong. Volume is how you find out what you are actually good at, and you cannot find that out from three songs.

What this looks like on paper {#on-paper}

Twelve months. Six to nine release dates written down now. For each one, a song chosen by what twenty people did rather than what you feel, and a check on plays-per-listener four weeks later that tells you which direction to lean next time.

Your move. Open your dashboard and calculate streams divided by listeners for everything you have out. The song at the top is probably not the one you would have guessed, and it is the one to build the next campaign around.