On this page

How Amplitude computes conversions through funnels

Identify key differences between Funnel and Event Segmentation charts.

When you calculate conversion for a funnel in which users can complete the steps more than once, Amplitude buckets each user based on the values tied to the first occurrence of each event.

Understanding these implications is vital to drawing accurate conclusions from your analyses.

First-touch attribution scenarios

Assume you have a funnel that tracks registrations (complete registration) broken down by the landing page each user sees (view landing page). If you hold constant by session_id, users must complete both steps of the conversion process in the same session for Amplitude to count them as converted.

Scenario 1: Funnels using both Hold property constant and group conversions by

When a funnel analysis uses both the hold constant and broken down by functions, Amplitude bases conversion on the earliest entry within the session.

Consider the following events and conversion results as examples:

When you hold constant by session ID, your chart displays the number of user sessions that included a conversion, and not the number of users who converted.

Scenario 2: Funnels using group conversions by, but not Hold property constant

When only the broken down by function applies, Amplitude bases conversion on the earliest entry within the lookback window.

Amplitude groups users by the first landing page they saw within the lookback window (how they entered the funnel). Amplitude considers them converted if they trigger the final event within the duration of the conversion window.

For example, consider the following events and related conversions:

Amplitude calculates conversions broken down by a filter differently than a group-by. A group-by looks for the earliest and most complete conversion first, and then groups by the specified property value. A filter-by first accounts for the property you filter by before looking for a conversion.

The logic of unique user counts in funnel analyses

When counting by unique users, the baseline conditions for conversion are:

  • A user must qualify for inclusion in the funnel: Amplitude can't filter the user out through the user segmentation panel. Any filters set in the Segment by module only apply at the time the user triggered the first funnel event.
  • A user must enter the funnel and complete all steps in the conversion window: The user must enter the funnel and complete all steps within the stated conversion window to count as converted in the final funnel step. Otherwise, Amplitude counts the user based on how far they progressed through the funnel.

When Amplitude counts by uniques, it only counts the earliest and longest conversion for each unique user:

  • Longest: In this context, longest means the most complete conversion—the completion of the most required steps within the funnel.

If Amplitude finds multiple conversions that meet the longest definition, it selects the first one and counts that as when conversion occurred.

  • Earliest: Amplitude measures earliest using the first converting sequence chronologically, if more than one exists.

When you use the broken down by function, Amplitude continues to use the longest/earliest logic to bucket users according to the property that was present at their point of funnel entry.

If you use Hold constant by in your analysis, Amplitude looks for the longest/earliest converting sequence within the same user session. When it does, the unit of measurement changes to unique user and session ID pairings.

When counting by event totals, the earliest/longest logic doesn't apply. Instead, Amplitude considers all conversion paths taken or attempted, rather than just the earliest/longest path per user. Amplitude then attributes the paths to the property for the event in the step it was broken down by.

Funnels versus event segmentation

The Funnel and Event Segmentation charts provide different types of analyses, and can display different results. The following table highlights some of those differences:

Was this helpful?