The complaint is now so familiar that it has lost most of its force: recommendation systems reward the middle, punish the difficult opening, and produce a culture optimised for retention rather than for anything else. All of this is broadly true and has been argued well elsewhere. What is less examined is what a deliberate refusal actually costs, and who can afford to pay it.

The work being made outside these systems has a recognisable shape. It is longer, or slower, or begins badly on purpose. It is frequently expensive to produce relative to its audience. And it is almost always financed by something other than the audience it reaches — a subscription base, a patron, a second job, a back catalogue.

The economics of being found

Distribution used to be the hard part and discovery came free with it. That has reversed completely. Anyone can distribute; almost nobody can be found. The consequence is that the marketing budget and the production budget have swapped places in the accounts of a great many small cultural enterprises, and a good deal of what looks like artistic decline is a straightforward funding reallocation.

We spend more on being findable than we do on the thing being found. Everybody in this business knows it and nobody says it on a panel.

An independent publisher

The organisations handling this best have mostly stopped trying to win at discovery and started building direct relationships that do not require it — mailing lists, physical events, print objects, memberships. All of these are old. All of them are expensive per person reached. All of them are considerably more durable than a position in a feed.

  • A direct list of four thousand people outperforms an algorithmic reach of four hundred thousand for almost any transaction that matters.
  • Physical objects and live events are the two channels that cannot be de-ranked overnight.
  • The cost of the relationship is paid once; the cost of visibility is paid every time.

What survives

It would be neat to conclude that difficult work is having a resurgence. The honest position is more mixed. Some of it is, in specific niches with unusually committed audiences and low production costs. A great deal of it has simply stopped being made, and the people who would have made it have gone to do something else, which is a loss that does not appear in any dataset because it consists entirely of absences.

What has genuinely changed is the disappearance of the middle. The very large and the very small both have viable models. The mid-sized cultural organisation — too big to run on goodwill, too small to command attention — is the one currently disappearing, and it happens to be the size at which most careers used to begin.