Writing.
Long-form essays on SaaS positioning, brand strategy, AI search, and the gap between what homepages claim and what reviews show.
Companion to the weekly public audit series. Same evidence-led methodology, more space to wander.
-
What 5 AI tools say about Plausible Analytics: the gap between brand citation and brand discussion
The fourth weekly audit. Plausible won the privacy round; cookieless is now the floor for every entrant. Plausible has already shipped the next category claim (AI-source attribution, documented in a December 2024 blog post). Zero of 296 Reddit posts captured 22 May 2026 associate Plausible with the feature. High brand citation, low brand discussion. The gap is community uptake, not content.
Read the essay -
What five AI tools say about Notion (and the bet they don't yet name)
The third weekly audit. Five AI tools converge on what Notion is: block-based, all-in-one, flexibility-vs-complexity trade-off. None yet name Notion Agent (the homepage hero today). Plus the 4 May credit-pricing shift and 237 Reddit posts showing the Custom Agents pricing revolt the LLMs haven't surfaced.
Read the essay -
What five AI tools say about Carrd (and the question Carrd's data doesn't answer)
The second weekly audit. Five AI tools converge on Carrd's positioning, held with discipline since 2016. The same five tools don't name v0, Lovable, Bolt, or Framer-AI as Carrd's competitors. Two readings of that silence, both plausible from the public data.
Read the essay -
What five AI tools say about Linktree (and why their homepage doesn't say it back)
The first weekly audit. Five AI tools agree on Linktree's actual moat. Linktree's homepage doesn't claim it. The gap between AI-tool consensus, homepage claims, and what 49 verbatim reviews actually show.
Read the essay
Want one for your SaaS?
The full 3-day Rational Magic sprint. Evidence-backed positioning. Brand OS File you can drop into any AI tool. $1,497 USD. Founding rate $497 (first ten customers, code FOUNDING).
Start your sprint →