🧠 ConversionDNA · Machine Learning Sales Prediction

What makes your users buy?

ConversionDNA decodes raw customer behavior data into high-resolution binary signals to isolate the exact user attributes that drive your revenue.

Required CSV Data Structure

Upload any CSV with your user attributes and conversion outcomes:

  • user_id: Unique identifier for each user/row.
  • is_success: Binary target column (1 = Buyer / Converted / Upgraded, 0 = Free / Churned / Abandoned).
  • Attribute Columns (Unlimited): Categorical text columns (auto-encoded with One-Hot Encoding) or numerical columns (auto-split into 10 deciles).

How the AI Model Works

ConversionDNA trains an ensemble Random Forest Classifier in-memory. By testing thousands of non-linear decision splits, it calculates feature importance to show you which user characteristics have the strongest correlation with conversion.

Frequently Asked Questions (FAQ)

How many rows of data do I need?

ConversionDNA works with datasets starting from a few hundred rows, but 1,000+ rows provide higher statistical confidence and sharper feature splits.

Is any of my customer data saved?

No. Trustometer operates entirely client-side and in-memory. Data is analyzed immediately and wiped as soon as your session ends.