I signed a record last spring on instinct. Heard it once, knew within about ninety seconds, wrote to the artist that evening.
That decision took a minute and a half. Everything after it took four months.
Not four months of work on the music. Four months of small things: chasing a master through three rounds of revisions, getting artwork resized, checking metadata, remembering that the pre-save needed to exist before I could promote it, following up on a follow-up I had already forgotten twice. The part that made the release good was the ninety seconds. The part that made it happen at all was the four months.
Ask most people what a record label does and they will describe the ninety seconds. Finding the artist. Hearing it before anyone else does. Shaping a sound until it lands. That part is real, and it is the part I love. But it is not where my weeks go, and I do not think I am unusual in that.
The scarce resource is attention, not taste
Here is what surprised me when I started paying attention to my own time. The decisions that genuinely need a human, the ones only I can make, are rare. Which artist to sign. What a campaign should feel like. When to hold a release back because it is not ready. You make a handful of those a month, and they are exactly the parts of the job worth protecting.
Everything around them is coordination. Dozens of small, unglamorous tasks spread across several releases, all moving on different clocks. None of it is difficult. All of it is relentless, and there is one of me.
So the constraint on my label was never a shortage of good ideas. I have more music I want to put out than I have capacity to put out well. The constraint is that every release demands weeks of remembering, and remembering does not scale. That is the actual ceiling on how much a small label can do, and almost nobody names it, because "we are limited by admin" is a much less romantic story than "we are limited by vision."
Which is why the tools have always felt slightly wrong
Once you see the label job that way, a lot of music software starts to look like it is aimed at the wrong target.
Plenty of tools help you promote better, get onto more playlists, understand your streaming numbers, distribute more widely. Those are real jobs and there is good software for each. But they all sit on the assumption that your problem is doing more of a thing. My problem, most weeks, is not doing more. It is not dropping what I have already committed to.
That gap is why so many labels end up running their actual operations in a spreadsheet or a Notion doc held together with hope. Not because they could not afford software, but because the software that exists was not built for the part that hurts. The lifecycle of a release, the deliverables, the deadlines, the artist back-and-forth, has quietly been nobody's category.
Where this gets interesting
I did not fully appreciate what this reframe implies until I started building.
If the bottleneck were creativity, then artificial intelligence would be a threat to the heart of the job, and the anxious conversations in music right now would be pointed at the right thing. But if the bottleneck is coordination, the picture inverts. Because coordination, the remembering, the following through, the holding of a hundred threads at once, is precisely what humans are worst at and software is increasingly good at.
That is a much more hopeful framing than the one currently doing the rounds. It suggests the useful thing to build is not a machine that makes taste. It is a machine that carries the load underneath taste, so the people with taste have room to use it.
I want to be honest that I am still working out where the line sits. Some of what I call coordination has judgment tucked inside it. Deciding when to push an artist and when to leave them alone is not really an operational task, even though it looks like one on a checklist. Drawing that boundary carefully is, I think, the hard and interesting problem, and it is what the next few posts in this series are about.
But the starting point I keep returning to is this. The romantic version of running a label is taste. The actual version is follow-through. And if we build for the second, we get to keep more of the first.
That is what I am building Labelflow around. If you want the story of how I got here, it starts with the first post in this series.