AI Music Growth: What a Bot Scare Taught Me About Real Fans
Gary WhittakerAI Music Audience · Case Study
A streaming spike is not the same thing as an audience.
One of my earliest AI-music growth lessons came from Sanctuary. The numbers moved fast enough to feel exciting—and strange enough that I could not tell whether I had found real listeners, landed in unhealthy playlist traffic, or simply misunderstood what the data was telling me.
This is a first-person case study, not a claim that I proved what caused the unusual activity. The original 2025 article speculated too freely about bots, release days and algorithms. I am keeping the experience and separating what I observed from what I could actually know.
What happened with Sanctuary
I saw unusual streaming spikes around the track. Later, DistroKid notified me about unusual activity connected to Spotify and the track was affected there while remaining available elsewhere. That gave me a real problem: the headline numbers had grown, but I did not know how much of that activity represented people who had intentionally chosen to become listeners of my work.
At the time I considered several possibilities, including problematic playlist traffic and my own use of DistroKid discovery tools. Those were hypotheses, not established causes.
If a metric rises but you cannot identify any corresponding increase in recognizable human behavior—returns, saves, comments, replies, newsletter joins, direct visits, repeat listening, purchases or genuine conversations—treat the spike as a signal to investigate, not proof that an audience has formed.
The mistake: letting the dashboard change the creative plan
For a while I reacted to the apparent traction like I suddenly needed to operate as a label: fewer songs, tighter release cycles, more teasers, more attention to dashboards. Some of that discipline was useful. The mistake was letting uncertain numbers dictate the creative identity before I had enough evidence that real people were asking for more.
I stopped listening to my own catalog as a creator and started watching it primarily as data. That was backwards.
What I no longer assume about release timing
The original article treated Tuesday as a broadly superior release day because Friday is crowded. I would not present that as a rule now. Release timing should follow the campaign, platform requirements, pitching windows, your audience habits and the reason for the release.
If you are testing timing, test it as your own operating choice and compare results over multiple releases. Do not turn one schedule into universal algorithm advice.
What counts as a stronger audience signal?
| Signal | What it can tell you | What it cannot prove alone |
|---|---|---|
| Stream spike | Something changed in exposure or playback. | That loyal fans were created. |
| Follower increase | More accounts chose to stay connected. | That they will return or support the next release. |
| Saves / repeat listening | The track may be useful enough to revisit. | A durable relationship with the creator. |
| Replies / comments / DMs | A person cared enough to respond. | That every interaction will convert into a customer or fan. |
| Newsletter / owned-audience action | Someone chose a connection you can reach again outside the streaming feed. | Guaranteed future engagement or sales. |
Be your own first repeat listener
One of the better lessons from this period was personal: I need to remain immersed in the work instead of treating every song as a unit of content. If I do not want to revisit the music, understand the catalog and recognize how the songs connect, I cannot reasonably expect strangers to build that relationship for me.
Strategy should support the creative work, not replace the reason the work exists.
The sustainable audience question
Instead of asking “How do I make the algorithm push this?”, I now prefer questions I can act on:
- Who is this song actually for?
- What part of it gives that person a reason to respond?
- Where can they encounter the story, process or project around it?
- What is the next useful connection after the stream?
- Did the next release bring any of the same people back?
That last question matters. A real audience becomes more visible through return behavior, not one impressive screenshot.
Where to go next
Case-study note: this page records my experience and what I learned from it. It does not identify the technical cause of the original unusual streaming activity.
