Methodology & Limitations

How Modest Idea identifies which demographic segments experience the problems your product solves.

How It Works

  1. Your product description is analyzed to extract the core problem, solution, and target audience
  2. 200-250 synthetic personas are sampled from our Census-grounded database
  3. Each persona is evaluated for problem-solution fit using multiple AI models
  4. Results are clustered into 5-8 demographic segments ranked by fit
  5. Top and bottom segments receive detailed reasoning analysis
  6. 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