Audience Intel Mirror — audienceintel.app

Across the captured pages (/, /buy, /for-llms, /research, /diagnose), there is no team page, no founder biography, no partner photograph, no academic adviser named, no signed essay, no first-person partner voice. Cambray Design Limited is named as the producer; ian@cambraydesign.co.uk appears in the footer. Every archetype’s reading hits the same wall.

The Anxious First-Timer scores 18 / 100 on Desire (catastrophic, weighted 0.35) — the partner photograph is the single most over-weighted rubric item for this archetype, and there is none. The Referred Sceptic cannot complete the trust transfer from referrer to recommended individual because the recommended individual is absent. The Considered Researcher has no specific partner to write the referenced email to. The Sophisticated Re-Buyer reads single-founder positioning where they need depth-of-bench. The Time-Poor Pragmatist weights Desire only 0.10, but the absent named human still costs ~5 points on the headline.

Every homepage purchase or pricing CTA links to a #waitlist anchor. The /buy page presents a Stripe button that the launch-week roadmap confirms is non-functional during pilot (env vars + Supabase Storage + ZIP upload pending). There is no published phone number, no Calendly booking link, no “book a fifteen-minute call” route. The only direct contact path is a footer email address.

The Time-Poor Pragmatist scores 18 on Action (weighted 0.40), pulling the composite to 35; this is the single largest negative contribution in the entire run. Every other archetype loses 3–8 headline points to the same cause.

The diagnostic surprise: the site advertises a £27 product and tells the buyer they can have it “on your own LLM, indefinitely” — and then will not actually sell it to them. The page does the hardest thing right (price in the hero subhead, plain-English CTA copy on /buy) and the easiest thing wrong: a working button.

Three pages disagree about the most basic spec the firm publishes. Homepage Pricing tier says “six calibrated buyer archetypes ... archetypes 07–10 ship in v1.2”. /buy says “five archetypes”. /for-llms Specification table — Number of buyer archetypes row says “Five core archetypes ... yielding 25 vertical variants”. /for-llms Specification table — Output deliverable shape row says “× three buyer archetypes”. /for-llms Anti-pattern-match clause says “three calibrated buyer archetypes”.

The Considered Researcher reads this as a coherence-failure test the site has set for itself and is currently failing. The Referred Sceptic reads it as “I’d better check with Sarah which page she meant” and the contact moves out of the site’s hands and back into the referrer’s. The Sophisticated Re-Buyer’s auxiliary procurement-instinct reads it as a tell.

Status note. Commit 2be8f30 (one session before this self-test ran) reconciled the canonical copy in Sections.tsx and for-llms/page.tsx. The captured pages still reflect the pre-commit state because no deploy has shipped since. Deploy is the missing step; the copy is correct in the repo.

Two URLs are referenced from the homepage footnotes, the /research index, and the bibliography — /research/library and /research/studies — but neither page is built. The launch-week roadmap puts these in Phase 2 + 3 (late May), but the live site already cites them as the costly-signal evidence base.

The Considered Researcher, whose Desire stage rewards citation traceability above almost any other factor, follows the homepage’s reference 1 (“Lecinski 2011 — see library entry”) and hits a dead link. Confidence collapses. The Sophisticated Re-Buyer reads the absence as “no public completed work” and discounts the bibliography claim accordingly. The Anxious First-Timer reads “thirty verified sources, full citations, pull-quotes” promised, 404'd, and reads it as the site being in transition.

The interactive revenue slider in the hero produces “£120,000 / year ... a 4,444× ROI before any second-order effects”. Four of five archetypes penalise this explicitly.

The Anxious First-Timer reads 4,444× as “the register of a webinar I would close out of” — Damasio’s somatic marker is negative. Loses 7 points. The Sophisticated Re-Buyer reads ROI multiples as evidence the firm does not respect them. Pooling signal in Spence’s sense. Loses 10 points. The Considered Researcher reads it as a Cialdini-adjacent move and activates manipulation-detection heuristics. The Referred Sceptic notes the over-performance as inconsistent with the editorial restraint elsewhere on the page.

Only The Time-Poor Pragmatist registers it neutrally — they are scanning for price, find £27 in the hero subhead, and do not slow down for the multiple.

The /diagnose page asks the visitor: “Does your homepage’s H1 name who you serve, what they need, and the outcome — within nine words?” The audienceintel.app homepage H1 is “The 20% of product-market fit your competitors found and you haven’t.” — twelve words, naming no vertical, no service, no outcome.

This is The Time-Poor Pragmatist’s reading specifically. The other archetypes either weight Attention low or read past the H1 to the hero subhead, which does name the substrate. But the recursion — the firm selling a clarity diagnostic, and failing its own diagnostic — is the most quotable single finding in the report. The firm is selling clarity it has not yet bought from itself.

The hero status line says “Closed-pilot pricing locks 16 May 2026”. The waitlist section adds “open access pricing increases 30% at launch”. The Why Now section ends “cost of being late is structural”. The Familiar section opens with “you were removed before they reached your About page. You will never know” — shame register.

The Anxious First-Timer reads urgency markers as a Layer B red flag and bounces. The Considered Researcher reads “locks 16 May” as a manipulation cue inconsistent with the rest of the page’s editorial register. The Referred Sceptic notes the tonal mismatch with the warm peer-recommendation that brought them to the site.

Brand Pack v4 explicitly forbids urgency markers. The current site contradicts the brand pack’s own stance, which the methodology surfaces independently as a finding.