Lifecycle stages

Lifecycle stages lives in the People area of the dashboard, under Lifecycle.

At a glance#

What it does#

Lifecycle stages divide your users into an ordered progression — new, active, power, at-risk, churned — where each stage is a rule and the distribution across them is computed live by the segment engine.

Each stage takes a single rule, from three kinds:

• RFM bucket — recency, frequency and monetary value combined into a standard bucket.

• Recency — active, or not active, within N days.

• Event — has, or has not, performed a given event.

Stages are ordered and colour-coded, and a user falls into the first stage whose rule they match, so the order is part of the definition rather than presentation.

When to use it#

As the standing summary of where your user base sits. A single active-user number cannot tell you whether you have a growing population of committed users or a churning one of the same size — the distribution across stages can.

Track how the distribution moves, not the absolute counts. Movement between stages is the actual signal.

Workflow#

1. 1

   Define the stages in order, most-engaged first

   Because the first matching rule wins, a broad rule placed early swallows the stages after it.

2. 2

   Use recency for at-risk, RFM for value

   "Has not opened in 14 days" is a good at-risk rule and a bad power-user rule. RFM is the reverse.

3. 3

   Check the distribution adds up to your population

   A large remainder means your stages do not cover everyone, which is usually a missing catch-all at the end.

4. 4

   Watch the at-risk → churned flow

   That transition is the one you can still do something about. Once someone is churned the intervention window has closed.

Permissions and prerequisites#

Requires event ingestion, and identified users for the counts to describe people rather than devices.

Limits and edge cases#

• A fixed "idle for N days" rule treats every user the same. A monthly customer behaving monthly looks at-risk under a 7-day rule. If that matters, User health measures idleness against each user's own rhythm instead.

• Distribution is computed live, so it moves without anyone editing anything.

• A stage that errors reports as an error rather than as zero, so a broken rule is visible instead of silently emptying a stage.

Troubleshooting#

Everyone is in one stage. An early rule is too broad. Reorder, or tighten the rule above it.

A stage shows an error. Its rule failed to evaluate — usually an event name that no longer exists.

Where the data comes from#

Served by

• Lifecycle automation

• Growth scoring

Related#

• User health

• Segments

• User journeys

• People overview