I Could See the Numbers, and Within a Year They Were Choosing My Subjects

For about four years I could see, immediately, how many people read and responded to everything I posted. I did not chase the numbers. I would have told you at the time, sincerely, that I wrote what interested me. Looking at the archive now, the drift is unmistakable and it starts about eight months in.

What happened is not that I started writing for engagement. It is subtler and I think more common: my sense of what was interesting changed, in the direction of what got a response. A subject that had done well felt more promising the next time it came up. A subject that had landed flat felt, somehow, less like something worth writing about. At no point did I make a decision.

The evidence is in the shape of the archive. Over four years, three of my recurring subjects grew and four disappeared entirely, and the three that grew were the three with the highest median response. Two of the four that disappeared were things I still find genuinely interesting and had simply stopped thinking of.

What I find most uncomfortable about this is that it did not feel like influence. There was no moment of temptation to resist. The metric was present, continuously, in the same place I did the work, and over enough repetitions it shaped what came to mind, which is a level below where any resolve operates.

What I do now, having moved to writing in a place where I control the visibility of the numbers: I look at them once a month, on a specific day, and never otherwise. This is a crude fix and it works better than any amount of intention did. The mechanism seems to depend on immediacy — a number attached to a specific post shortly after posting is what does the shaping — and a monthly aggregate simply does not have the same grip.

The second thing is that I keep a list of subjects I want to write about, written when I am not looking at any numbers, and I take the next one from the list rather than from what occurs to me. What occurs to me is the compromised channel. The list is roughly eighteen months old and about a third of the items would never have surfaced on their own.

I want to be careful not to overstate this into an argument that audience feedback is bad. Some of the drift was toward things I was genuinely better at, and the responses were real information about what was useful to people, which is not a trivial thing to disregard. The problem is not the information. It is that the information arrived continuously, attached to my own sense of interest, in a form I could not separate out — and by the time I noticed, four of my subjects had gone quietly and I could not tell you when.