Compare Amplitude Funnel Segments
Compare a few relevant segments under the same funnel definition, then inspect who enters each group. Segmenting and grouping semantics matter: a property can be evaluated at funnel entry rather than at the final action.
Before you start: Have a validated funnel and properties whose timing and coverage are understood. Agree which difference would affect a product decision.
Worked example: Compare self-serve and invited workspace owners through first report creation. Keep the conversion window equal and confirm that acquisition_source is known when each user enters.
Step-by-step guide
State the hypothesis
Explain why the chosen segment may experience different friction. Avoid scanning every property until a dramatic chart appears.
Validate the segment field
Check actual event or user property values and historical coverage. Confirm when the property is evaluated in this funnel.
Keep the funnel fixed
Use identical steps, order, counting unit and conversion window. Change the segment, not the definition of success.
Inspect counts and drop-offs
Show entrants and completions for each group. A large percentage difference from a tiny segment is not strong evidence.
Review representative histories
Check users near the largest drop-off and a few converters. Look for different eligibility, role or instrumentation rather than assuming the interface caused the gap.
Document the interpretation
Save the query and unresolved confounders. Follow up with a targeted analysis or experiment before claiming causality.
Common mistakes
Properties assumed current
Funnel segmentation can depend on entry-time values.
Tiny segments overinterpreted
Show counts and uncertainty.
Changing multiple settings
Then the source of the difference becomes unclear.
Tips
Limit the initial comparison to a clear hypothesis.
Keep segment coverage dates visible.
Inspect both a converter and a non-converter.
Sources and verification
Based on product documentation checked on 2026-10-24. This is a documentation-based guide, not a hands-on benchmark. Features and availability can change.