Method
Reading funnel drop-offs without lying to yourself
A steep drop between step two and step three feels urgent. That urgency is useful only if the drop is real. In Conversion Path Analytics, the first job is to ask whether the chart measures the same humans across steps.
Check the denominator
Many funnels silently change who qualifies mid-path. If step two requires login and step three includes guests who bounced earlier, your “drop” is partly a definition change. Write the eligibility rule above the chart before you redesign anything.
Separate delay from departure
Users who pause overnight look like abandoners in a same-day session funnel. Stitch a cohort window that matches how your buyers actually decide. If your product is researched for a week, a 24-hour path will invent panic.
Sample the exits
Pull twenty recent users who stopped at the suspect step. Read their last screens, support tickets, and device notes. Patterns appear faster than another dashboard tile. Then — and only then — draft an intervention with a falsifiable expectation.
Path Signal Lab spends a full module on this interrogation habit. If your team skips it, you will keep shipping UI for ghost problems.