In plain terms — no computer science degree required
The visibility machine: one hub, four spokes
This page belongs to the Sunday edition “Being seen in 2026: the machine behind the visibility”. That text covers the why — this one covers the how: a single drawing that shows how the five parts — Owned, Social, Found, Proof, Feedback — mesh into one closed loop. Ten minutes, no expertise required.
The drawing
The hub and the four spokes
Owned is the hub at the centre: website, blog, newsletter, RSS feed — the surfaces the company itself owns. Everything else in the drawing is borrowed; the hub is not. That is why every spoke points back to it: whoever arrives via LinkedIn, a search engine or an AI assistant ends up on your own ground — your own address list, your own domain, your own archive.
Social and Found are the two routes by which people find the hub. Social is deliberately borrowed reach: LinkedIn, because that is where the decision-makers are; Bluesky, because there your own domain is your name; a self-run Fediverse node, because ownership in social media does not get more radical than that. Found is being found, redefined: the old search engine craft, plus the new discipline of getting cited by AI assistants — which cite pages with clear claims, sources and authorship.
Proof and Feedback keep the machine honest. Proof means: the work itself lies open — decisions, failures, code — and thereby proves what references merely assert. Feedback means: cookie-less, self-hosted counting plus a tag on every link, so the machine does not run blind — and still nobody gets surveilled.
What travels where?
Out of the hub flows content (light blue): one teaser per blog post to the social channels, scheduled in advance; citable pages towards search and AI assistants; the openly documented work as evidence. Back flowsattention (blue): readers who follow the teaser; readers who arrive via a citation; trust from those who have checked for themselves.
And at the bottom left, measurement closes the loop (amber): every link carries a tag, visits are counted without cookies, and in the morning an agent pulls the numbers — which channel brought readers yesterday, which ones stayed? Those insights flow back into the hub and help decide what gets written next week.
The drawing also marks what is machine and what stays human: scheduling teasers and pulling numbers is automated. Replying and writing are not — deliberately.
Staying honest: what the machine cannot do
Three things need saying, or this drawing would be advertising rather than explanation:
- It does not run without human fuel. Replies come by hand, trust is built in conversation, and a community cannot be automated. An account that answers automatically stops being credible — and, incidentally, breaks the rules of some platforms.
- It costs real hours. The build was roughly one gross working week, scattered over five weeks; the operation costs server francs in the low double digits per month — and two to four hours of writing every week. That is the one line item that never shrinks.
- Reach compounds slowly. Your own list grows reader by reader, not in leaps; the compound interest takes months. Whoever wants fast numbers rents them on a platform — and can then read, in the red line of the drawing, who that reach belongs to.
You do not have to believe this — you can check it
The nice thing about this machine: it stands in public. Website, blog, feed, the channels, the missing cookie wall — all of it can be verified from the outside, without asking anyone; that is exactly the part called “Proof”. And the drawing works as a checklist for your own company: which of the five parts is missing entirely for you? And if the answer is “Owned” — how much of your reach do you actually still own?
What this machine costs to build and to run — and why, without AI agents, it would have remained a project for an agency — is in the Sunday edition: “Being seen in 2026: the machine behind the visibility”