Upload a CSV, pick your yes/no outcome and up to 8 candidate drivers, and get a full logistic regression — odds ratios with confidence intervals, ranked driver influence, and honest model-quality metrics. Free.
or click to browse · max 3 MB
Free analyses run on up to 10,000 rows. Larger files are randomly sampled to that size — sign up to analyze your full dataset.
Fitting logistic regression and computing odds ratios...
Sent to — odds ratios with confidence intervals, driver ranking, predicted-probability distribution, model quality metrics, R code, and AI insights.
Analyze another fileThe analysis fits a logistic regression: it models the log-odds of your yes/no outcome as a function of the drivers you mapped. Each driver's effect is reported as an odds ratio — how much one unit (or one category) multiplies the odds of the outcome — with a 95% confidence interval and significance test. Drivers are ranked by the size of their z-statistic, model quality is summarized with AUC, McFadden pseudo-R-squared, and confusion metrics at a 0.5 threshold, and the predicted-probability histogram shows how cleanly the model separates the two classes.
Use it whenever the thing you care about is a yes/no — churned, converted, clicked, defaulted, passed — and you want to know which measurable factors move it and by how much.
Not for numeric outcomes (use linear regression), outcomes with three or more categories, or time-to-event questions (use survival analysis).
Built for: Growth, retention, risk, and operations analysts explaining binary outcomes
Typical data source: A customer or transaction table with a yes/no flag column and several attribute columns
One row per customer with a churn flag and candidate drivers:
Minimum 30 rows · Best with 200-50,000 rows with a 10-90% outcome rate
Standard-library analysis: which factors drive a binary outcome — churned or retained, converted or not, defaulted or paid? Logistic regression turns each driver into an odds ratio you can act on, ranks the drivers by statistical influence, shows how cleanly the model separates the two groups, and reports honest quality metrics (AUC, accuracy, sensitivity, specificity). Works on any dataset: map a yes/no outcome and 1-8 candidate drivers.
Each driver's multiplicative effect on the odds of the outcome, with confidence intervals — above 1 raises the odds, below 1 lowers them.
Drivers ranked by statistical influence so you know where to act first, each tagged with its direction.
How confidently the model scores each row — two separated humps mean the drivers really split yes from no.
AUC, pseudo-R-squared, and accuracy versus the majority-class baseline — an honest read on how much signal the drivers carry.
Plain-English interpretation — what the numbers mean, what's significant, and what to do next.
What actually drives my churn?
Map your churned/retained flag and the customer attributes you suspect matter. Each driver comes back as an odds ratio — 'every $10 of price multiplies the churn odds by 1.6' — ranked by statistical influence, with honest model-quality metrics so you know how much to trust it.
See our FAQ for details on pricing, data privacy, and how the analysis works. Every report includes a Methodology section showing the statistical test, assumptions checked, and diagnostics run.
Run any analysis on your own data — validated R analyses, interactive reports, AI insights, and PDF export.
Try Free — No Credit Card