Evidence records5 customer · 1 method
Customer decisions and research checks.
What customers tested, what the models estimated, and which outcomes were observed or attested.
Customer decisions
Nº 06FintaFinta reported a 2.73× change in its annual subscription price, from $348 to $950, and 5 new customers on its reintroduced monthly plan.Synthetic Respondents and Causal AIRead case study →2.73×Nº 01A commercial spaceflight leaderA modeled space tourism offering for a segment of fewer than 30,000 people; $500M in avoided research cost is a Subconscious founder-attested estimate.Synthetic Respondents and Causal AI — three experimentsRead case study →$500MNº 02River City Federal Credit UnionRiver City Federal Credit Union reported a 42% increase in auto-loan ad Result Rate over its prior campaign, from 0.52 to 0.74.Discrete choice survey with digital twins, plus Facebook/Instagram A/B testingRead case study →42%Nº 03MintelSynthetic respondents showed 0.79–0.88 Jensen-Shannon distribution similarity from demographics alone, with no choice data as input. This is distribution agreement, not predictive accuracy.Choice-based conjoint replication; hierarchical Bayes; Jensen-Shannon divergenceRead case study →0.79–0.88Nº 04New Age FloralThe simulator favored a $60 OshunPouch® price and home-compostability wording; the record does not document a post-study sales result.Causal Research and Market Simulations — five iterative AI-driven experimentsRead case study →$60
Method validation
Parameter-rank evidence from 43 design-filtered human studies. The paper reports limits and misses. Read the paper.