KD
Drivers & lift
Key Driver Analysis
Upload a survey export, pick your outcome, and see which levers move it the most. Everything runs in the browser - no uploads beyond this page.
Client-side | ridge/logit
Load data Awaiting file

Excel or CSV with headers in row 1. Blank rows are ignored.

No file loaded.
Outcome Auto-detect
Binary outcomes use logistic regression; numeric scales use ridge regression.
Drivers

Mix categorical and numeric predictors. Very high-cardinality text fields are skipped.

Load a file to see fields.
Settings Stable defaults
Controls shrinkage to keep coefficients stable. 0 = none, 1 = strong.
Gradient steps for the logistic model. More steps for noisier data.
Load a file to get started.
Upload data, select an outcome and drivers, then run the model.