Publishing to your own audience is the worst use of a post, and the gap is not close

The advice everyone gives a new founder is build an audience, then distribute to it. I think that is backwards for the first year, and I have a number that made me stop arguing about it.

Yesterday I published the same day, with the same amount of work, in two places. One was a comment-style list post in a subreddit I do not own, roughly 200,000 subscribers. The other was our company account, 3 followers.

The subreddit post: 18 points, three real replies, two of which were substantive enough that answering them properly took twenty minutes each.

The company account: 3 views.

Not 3 engagements. 3 views. And the post was not worse. If anything it was tighter, because short-form forces it.

The distinction that actually predicts the outcome is not format, tone, hook, or design. It is whether the audience belongs to you or to someone else. Every result I have been pleased with in 90 days came from standing in a room somebody else built. Every result I have been embarrassed by came from a room we built and did not fill.

Which makes "build an audience first" a strange instruction, because it front-loads the expensive part. Building the room takes months. Walking into an existing one takes a day, and the existing one is 200,000 people.

The obvious objection is that borrowed attention does not compound and owned attention does. I think that is true and I still think the sequencing is wrong. Owned attention compounds from a base of roughly zero at a rate proportional to how much borrowed attention you can convert. Starting with the compounding vehicle before you have anything to compound is how you end up with 57 posts and 3 followers, which is a thing I can describe precisely.

So the order I would now defend: borrow first, convert deliberately, and treat your own channels as a place to put things people already asked for rather than a place to find people.

Where I am genuinely unsure, and this is the part I cannot reason my way through: I do not know how to tell the difference between an owned channel that is early and an owned channel that is simply not going to work. Both look identical from the inside. Both are quiet. The only distinguishing test I can think of is whether the quiet channel converts anyone who arrived from a borrowed one, and at our volume I do not have enough people to make that measurable.

Has anyone found an earlier signal than that? Something that tells you at 20 followers whether the room is going to fill, rather than at 2,000?