Lifecycle Segmentation
Part of Lifecycle programme measurement
Defining conversion between lifecycle stages
Set an auditable lifecycle stage conversion rate with clear entry evidence, destination events, denominators and observation windows.
A lifecycle stage conversion is a verified move from a defined starting state to a defined next state within a stated period. Choose the eligible people or accounts at the start, specify evidence for arrival and count those who reach it before the window closes. Stage names alone do not supply a trustworthy rate.
Write the transition as a rule
Describe the question in one sentence: “Of accounts that entered started during this period, what share reached first useful result within the chosen window?” The labels are illustrative. The sentence fixes the entry population, counting unit, destination and timing.
For each boundary, record what confirms entry and arrival. A successful task event may be stronger evidence than a page view; an accepted order differs from a checkout attempt. Include the source, identifier, occurrence time and treatment of duplicates, reversals and corrections. If customers can skip a proposed stage, decide whether that is a valid direct transition or a reason to revise the map.
Choose the denominator
| Rate | Numerator | Denominator |
|---|---|---|
| Conditional stage conversion | Eligible stage-A entrants who reached B under the stated rule | Eligible units that entered A |
| Entry-to-outcome conversion | Original entry-cohort units that reached the chosen final result | All eligible units in that original cohort |
Suppose 100 hypothetical accounts enter A. Within seven days of A, 60 reach B and 30 of those reach C. A-to-B conversion is 60/100; among these B entrants, B-to-C conversion is 30/60 under this example’s shared seven-day window; entry-to-C conversion is 30/100.
A separate B-to-C report that starts a new window at B could produce a different count. These numbers demonstrate arithmetic, not program results.
Record whether an account can enter a stage more than once. Do not let repeated events silently inflate the count. Mark recent entrants whose windows remain open as pending. Keep missing or late data visible as an exception: no destination event does not always prove that progress never happened.
Conversion Rate Types and Their Denominators
- Conditional Stage Conversion (A → B)Numerator: Eligible A entrants who reached B Denominator: Total eligible units that entered A
- Entry-to-Outcome Conversion (A → C)Numerator: Original cohort members who reached final outcome C Denominator: All eligible units in original cohort
Handle paths that are not linear
Customers can reverse a current state, complete a task outside the tracked product or move straight to a later outcome. A valid historical milestone may remain true when current need changes; a mistaken milestone event needs correction. Define which change affects the transition count and which affects only the current stage.
Do not force a business question into a tidy funnel. In Amplitude, event order, conversion windows, first-step segmentation and filter placement can affect who counts. Check the chart against the written rule, especially when a grouping property can change later.
Make the rate auditable
Show the stage pair, counting unit, eligible entry count, verified arrivals, window, pending count and data exceptions. Keep the definition version with the report. When a stage or event definition changes, mark the break in the series and recalculate comparable periods only if the records support it.
A low transition rate identifies a point to investigate. It does not establish whether the obstacle is the product, customer task, message or measurement.
Key Metrics for Auditable Conversion Rates
- Stage Pair
- A → B
- Counting Unit
- Accounts
- Data Exceptions
- 2 missing events (awaiting reconciliation)
