ssr-market-researchlisted

# SSR Market Research Estimate how a target market will react to product concepts by simulating consumers with an LLM and scoring their reactions with **Semantic Similarity Rating (SSR)**. SSR avoids the "everything is a 3" failure mode of asking an LLM for numeric ratings directly. ## Why direct numeric ratings fail (and what SSR does instead) If you ask an LLM "rate your purchase intent 1–5," it **regresses to the center** — low variance, unrealistic, ~50% accuracy. SSR fixes this with four design choices, all of which matter (dropping persona detail alone cut accuracy 92% → 50% in the paper): 1. **Rich persona conditioning.** Each synthetic respondent impersonates a *specific* person with demographics + backstory + context, not a generic "user." 2. **Free-text elicitation first.** NEVER ask for a number. Ask an open question ("How likely would you be to adopt/pay for this, and why?") and capture the natural-language answer. The variance lives in the words. 3. **Embedding → Likert mapping.** Embed the free-text reply and 5 anchor statements (one per Likert point). Cosine-similarity, subtract the minimum, normalize → a probability distribution over 1–5. 4. **Average over multiple anchor sets + many respondents.** 6 stylistic anchor sets capture different ways people express intent; many respondents give a survey-level distribution. The paper hit ~90% of human test–retest reliability this way, beating supervised models. ## The pipeline ``` segments +