Five skeptical-founder objections, answered with the artifacts that already ship.

The list is the format, not the product. Each issue names two picks and spends the next several pages on documented demand proof, unit economics back-solved from $1M, comparable exits with disclosed dates, a Confidence Margin score, and a falsification list — the operational assumptions the founder has to test before they commit. Read issue 001 in full to see the artifact, then open the archive to filter the same body of work by vertical and by Confidence Margin. If a pick scans as a one-paragraph hook, it is not this product.

Those formats trade depth for cadence: a tight hook, a thumbnail, a CTA, no recorded objections. Marginalia publishes a multi-page research packet every Friday, in one format, with the same five sections every issue. The difference is in what is sourced and what is falsifiable — every comparable exit and demand cite is dated and listed; every pick has a Confidence Margin and a set of assumptions the founder is expected to disprove. The published side-by-side scorecard against the most-circulated listicle format spells this out axis by axis.

Free issues ship once a week — they are the marketing surface and pull the most defensible pick from the Friday packet. The full packet (both picks, full unit economics rows, comparable exits, source disclosure, falsification targets) ships to paid subscribers at the back end. The two plans exist so a solo reader can pay monthly and stay alive to the cadence without committing to a year; the annual plan pays for exactly the same editorial work, billed once. The pricing page spells out what ships on each plan and what does not.

The archive exists precisely to filter the body of work before you subscribe — every published issue is browsable by vertical and by Confidence Margin, so a pick below your personal threshold can be skipped on the page, not on your invoice. The Validation Lab cohorts go further: small quarterly groups pressure-test picks in public and publish what breaks, including the picks where the lab disproves the original thesis. Both surfaces pre-date the question; they are the answer to it.

Solo founders and small teams (one to three people) building toward $1M ARR, who have already shipped something and need the next defensible idea rather than the first. It is not a feed for someone choosing a career path, not a get-rich format, and not a 100-startup-a-day stream. Read the sample issue to see whether the rigor reads like yours, then open one archived idea — pick any slug in the archive index — to gauge how the format reads across verticals before you subscribe.