I started writing in public. Here's what the numbers say.

I started writing at the end of August — August 27, 2026, to be exact.

Not to build an audience. To force myself to understand things properly. The fastest way I know to find the holes in what I think I know is to try to explain it to someone else: the moment you write "and therefore" and can't finish the sentence, you've found the thing you didn't actually understand.

So far that's 11 posts, 1.6K impressions and 34 reactions. This post is the receipt. Every number in it reads live from Buffer's API on the server, refreshed once a day — so it will contradict me the moment the data does.

11 posts is not a sample size. It's a start. But the patterns are already sharp enough to have changed how I publish.

The network matters more than the post

Every piece goes to both networks the same day, minutes apart — sometimes word for word, more often rewritten to fit each one. Whatever separates these bars, it isn't the idea.

LinkedIn shows my posts to roughly 5× more people than X does. Same week, same argument, same person writing it. The distribution is doing more work than the prose.

Reach and engagement pull in opposite directions

The obvious read is that one network is simply better. The numbers say something more useful.

The network that reaches the fewest people — X — is the one where the people who do see a post are about 3× more likely to reply, react, or click. A big room where nobody speaks, and a small one where everybody does.

Read the right-hand side of that chart with suspicion, though. Those engagement rates come from denominators in the dozens. One reply on a post seen 24 times is 8%, and it means far less than the percentage suggests. The shape is real; the precision is not.

My most-seen post isn't writing at all

Ranked by reach, the post at the top isn't an essay. It's a hiring announcement from October 17, 2025.

On its own it was seen 1.4K times — more than every piece of writing since, combined. It's in a different colour above because it's a different kind of post, and dropping it would have made the chart flatter and less true.

That's the part worth sitting with. Reach follows what people need right now — a job, a link, a favour — not how carefully something was written. Which is a decent argument for writing to understand rather than writing to be seen: the first goal I control completely, and it pays out whether or not the second one ever does.

Still early

I'm keeping the cadence and letting this post update itself. It reads the API on every rebuild, so these charts will get less flattering or more interesting without me touching a word of it.

Come back in a few months. It'll be a different post.