The site started as a decision about whether to build at all, versus just posting to LinkedIn and calling it done. The case that won: a newsletter needs a home that isn’t rented, and four unrelated interests — legal tech, running, the workshop, and field notes — are worth one address instead of four scattered profiles.
The architecture followed from that. One repo, one Astro site, four pillar collections acting as rooms in the same house, not four separate sites pretending to be related. Subdomains were the tempting first instinct, and got rejected for three concrete reasons: they split search authority across four weak sites instead of one that compounds, they quadruple maintenance, and they contradict the premise — if the brand is “the mix is the point,” walling the audiences off from each other is architecture arguing with strategy.
Even the plumbing had to agree with that premise. The repo went under a new, separate GitHub account rather than the one already tied to Affinity/Lawyerist work — a home base that’s supposed to outlive any single employer shouldn’t be hosted under one.
Main takeaways
- “The mix is the brand” — one person, four subjects, beats four separate niche profiles.
- Subdomains looked tidy but split search authority, quadruple maintenance, and argue with the whole premise; path-based sections under one roof won instead.
- Even hosting decisions should agree with the site’s premise of outliving any single employer.
What I got wrong
- Didn’t budget enough early time for the plumbing — repo, hosting — before wanting to jump straight to design.
- No habit yet for keeping two GitHub accounts straight, which is a mistake waiting to happen the first time the wrong one is active.
