The decision
Funnel numbers are meaningful only when each step has a clear definition and consistent scope. Decide which hostname, date range, time zone, identity, and event define the step. A checkout-button click is not a checkout session, and a success-page view is not a verified purchase.
A worked example
A useful digital-kit funnel tracks public page visits, sample use, confirmed signup, checkout requested, checkout session created, verified purchase, download delivered, and refund. Anonymous browser identity can be carried into checkout metadata when analytics consent allows it. Purchases without that identity still count in the payment ledger.
How to put it into practice
- Write the authoritative source for each metric: browser events, subscription database, or payment provider.
- Use a fixed complete-day date range and exclude internal or preview traffic consistently.
- Monitor missing events and consent-related gaps before interpreting conversion rates.
- Compare payment records with analytics rather than treating analytics as the financial ledger.
A failure to plan for
An untracked purchase can look like zero conversion, while an unverified success-page event can invent revenue. Reconcile sources and label measurement gaps before making product or acquisition decisions.
Try it on your project
Define seven funnel steps in a table with event, source, denominator, and exclusions. Trace one sandbox purchase through the system and confirm that the purchase event can be linked to the intended checkout without using email as a public identifier.
Keep the next step small
Use the free demand scorecard or planning tools to make your assumptions explicit. The $19 launch kit brings the blueprint and seven editable worksheets together.