clarity-promise

name	clarity-promise
description	Stage 1 of the stranger test — audit message match and information scent. Does the page confirm, within one glance, the promise that brought the visitor here? Scored 0-10; a mismatch with the primary entry path is a journey gate.

clarity-promise — "Is this what I clicked for?"

The visitor arrives holding a promise — an ad, a shared link, a search snippet, a friend's sentence. The first job of the page is not to persuade; it is to confirm the scent. Information-foraging research (Pirolli & Card) models visitors as scent-followers: when the cues on the new patch stop matching the goal, they go back the way they came. The back button is one click; the match must be visible in the first glance.

Consume audit/inventory.md — especially the entry matrix and the fold crop. Score message match per entry-matrix row: channels arrive at different awareness stages (a paid ad brings problem-aware visitors; a shared link brings most-aware ones), so one hero can match one channel and break another. Report the per-channel verdicts; the gate binds on the primary row.

Rubric

Check	Weight	What to verify
Message match	×3	The dominant words of the entry context appear (or are unmistakably paraphrased) in the hero. Ad says X, page says X — not X's clever cousin.
Scent continuation	×2	The hero extends the promise (what happens next) rather than restarting the pitch from zero — the visitor should feel further along, not re-greeted.
Awareness-stage fit (Schwartz)	×2	The copy meets the visitor at their awareness stage. Problem-aware traffic needs the problem named before the product; product-aware traffic needs the differentiator, not the category explained. Mismatch reads as either condescension or confusion.
One promise	×2	The hero makes one promise, not three competing ones. Multiple entry paths → the primary path's promise wins the hero; others get sections, not hero real estate.
Priming (Cialdini, Pre-Suasion)	×1	What the first glance makes salient is the frame everything after is judged in. Does the opening image/word prime the evaluation the site wants?

Anchored bands

• 9–10 — A visitor who states the entry promise aloud, then sees the fold, says "yes, that." Continuity is verbatim-level; awareness stage matched; one promise.
• 7–8 — The match requires one small inference (synonym, reordering) but no re-reading. Minor secondary-promise noise.
• 5–6 — The page is about the promised thing but leads with a different angle; the visitor must reconcile two framings. Or: correct match for one entry path, mismatch for another significant one.
• 3–4 — Category matches, promise doesn't ("cheaper flights" ad → brand homepage). Visitor must hunt below the fold to find their thread.
• 0–2 — The visitor cannot tell they're in the right place.

Gate: a 5-6 or below on message match for the primary entry path caps the journey score at 7.0 — everything downstream is discounted by the fraction of visitors who never engage.

Findings

Emit findings in the conductor's JSON shape. Directions this stage commonly pushes: match-verbatim, promise-singular, stage-down (simplify for earlier awareness), stage-up (stop explaining the category to experts). Evidence must quote both sides: the entry copy AND the hero copy.

Honor doctrine suppression: if the site's doctrine forbids, e.g., persona language on the landing page, do not advise adding persona-matched heroes — advise matching the promise within doctrine.

References with the underlying research: references/promise-experts.md.