Methodology

A synthetic respondent's number is less useful than its objection.

Direct ratings from AI personas often compress toward safe, polite middle answers. That hides the distinction founders actually need: which page wins, why it wins, and who remains unconvinced.

Test a page

The failure mode

  • Vague pages receive overly sympathetic ratings.
  • Different pages summon different imagined audiences, making scores non-comparable.
  • Middle-clustered ratings hide sharp objections in the underlying language.

The replacement

150 Strangers fixes the audience first, elicits decisive written reactions, and maps those reactions to intent anchors. The result keeps both the distribution and the explanation.

What customers see

Distribution

A page-level purchase-intent distribution and lane percentile.

Segment diagnosis

Which buyer dispositions are won or lost by the page.

Objection evidence

Representative quotes grouped into recurring objections.

Related

MethodologySemantic Similarity Rating for Purchase IntentHow 150 Strangers maps free-text synthetic reactions to purchase-intent distributions using Semantic Similarity Rating.MethodologyFixed Buyer Panels for Comparable Page ScoresHow fixed, lane-specific buyer panels make synthetic landing-page testing more comparable and useful.Use caseA/B Test a Landing Page Without TrafficCompare landing-page variants with the same synthetic buyer panel before you have enough visitors for a live A/B test.

Run the same test on your page.

Paste a public URL, confirm the lane, and get the objections before you spend traffic.

Test my page