Methodology & Limitations
How Modest Idea identifies which demographic segments experience the problems your product solves.
How It Works
- Your product description is analyzed to extract the core problem, solution, and target audience
- 200-250 synthetic personas are sampled from our Census-grounded database
- Each persona is evaluated for problem-solution fit using multiple AI models
- Results are clustered into 5-8 demographic segments ranked by fit
- Top and bottom segments receive detailed reasoning analysis
- Final synthesis produces actionable findings and recommendations
What is PSF Score?
The Problem-Solution Fit (PSF) score measures how strongly a demographic segment recognizes and experiences the problem your product addresses. It ranges from 0-100.
- 70-100 — High fit: strong problem recognition
- 40-69 — Medium fit: partial recognition
- 20-39 — Low fit: minimal recognition
- 0-19 — Minimal fit: unlikely to relate
Where the Data Comes From
Our personas are grounded in the US Census Bureau's Public Use Microdata Sample (PUMS), part of the American Community Survey.
- 10,000+ statistically representative personas
- Real demographic distributions (age, income, education, occupation, geography)
- Weighted sampling preserves population proportions
- Each persona enriched with psychological and behavioral traits
Three-Layer Bias Mitigation
Synthetic persona evaluations can suffer from systematic bias. We use three independent layers to mitigate this:
- IPF Weights — Iterative Proportional Fitting ensures our persona sample matches real demographic distributions.
- Temperature Variation — Variable AI evaluation temperatures prevent herd mentality.
- Multi-Model Ensemble — Different model biases average out, reducing systematic error.
Important Limitations
- Results show who HAS THE PROBLEM, not who will PAY
- AI cannot simulate emotional, social, or irrational buying factors
- Reflects US population demographics only
- Synthetic personas may miss cultural/contextual nuances