You are looking at conversion through a signup funnel and want to separate signal from confounding. Several user attributes and product exposures move together, so raw correlations are not enough.
How would you use a simple regression to understand what drives conversion?
Choosing regression for a binary conversion outcomeInterpreting coefficients and odds ratiosSeparating raw correlation from conditional associationExplaining why regression alone does not prove causality