Methodology
Semantic Similarity Rating turns reactions into comparable purchase-intent distributions.
Directly asking an AI persona for a 1-5 rating tends to produce polite middle answers. 150 Strangers instead asks for natural reactions, then maps those reactions to calibrated intent anchors by meaning.
Test a pageThe scoring path
- A fixed persona reacts to the page in free text.
- The reaction is embedded with the same embedding model used for anchor statements.
- Semantic similarity to each anchor rank creates a probability distribution over intent levels.
- Persona-level distributions aggregate into segment means, page means, and peer percentiles.
Why free text matters
Free-text elicitation preserves the reasons behind a reaction. The report can show a purchase-intent distribution and the words that drove it: trust gaps, unclear benefits, missing proof, pricing hesitation, or category confusion.
What the score means
A 150 Strangers score is a comparative research signal. It is most useful when the same buyer panel judges two variants, a competitor set, or a lane-specific peer baseline.
Questions
Common questions
Does SSR predict conversion rate?
No. It estimates relative purchase intent and objection patterns from synthetic reactions. Conversion rates still depend on traffic source, price, timing, product quality, and many factors outside the page.
Why not use a normal Likert survey?
The product can show a Likert-like distribution, but it derives that distribution from meaning rather than directly asking synthetic respondents to choose a number.
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