User AnalysisChurned User Identification and Profiling
Establish the filter criteria and profile characteristics for churned users, ready to use directly in later analysis and win-back operations.
Segmentation
by clusterChurned User Identification and Profiling is a churn-definition and profiling capability in ThinkingAI's industry Skill library. It helps operations teams establish churned-user filter criteria (how long inactive counts as churned? which users are churning?) and describe churned-user profile characteristics. This is the prerequisite for all churn analysis and win-back operations, and if the definition is off, everything downstream is too. This is ThinkingAI's decade of industry experience, packaged into ready-to-use Skills.
Live walkthrough
#reg_time within the last 30 days + no login event in the last 7 days)All three are done: the definition recommendation, the cohort asset, and the persona characteristics. Conclusion: use 7 days; the 3-day camp would wrongly hit 18% of your normal players.
1. Why 7 days
| Days inactive | 30-day natural return rate | Judgment |
|---|---|---|
| 3 days | 18.3% | Nearly 1/5 come back on their own; too early to call it churn |
| 5 days | 9.1% | Still meaningful returns |
| 7 days | 4.2% | Return probability is near zero, can be judged churn |
SLG combat pacing is inherently intermittent, with "play a few days, rest a few days", so a 3-day definition would sweep many normal intermittent players into the churn group; whatever follows, analysis or win-back pushes, would be polluted by this batch of "fake churn".
2. Cohort asset created
#reg_time within the last 30 days AND no login event in the last 7 days3. Churned-user persona: four high-incidence traits
| Dimension | Share among churned users | Share among active users | Difference |
|---|---|---|---|
| From channel B | 45% | 20% | 2.3× |
| Android device | 68% | 51% | Churn about 2× that of iOS |
| Not in an alliance | 81% | 34% | Non-alliance players churn 3× |
| Level ≤15 | 72% | 28% | Concentrated in the early growth phase |
The most notable is alliance participation: 81% of churned users never joined an alliance. Combined with the level distribution (72% stall before level 15), the churn window most likely sits in the gap where "the solo experience ends but social bonds aren't yet formed".
On your data
That was a simulated run
Leave your work email and we will run a live walkthrough on your real business data.
The problem
Every team uses the term "churned user," but the definitions differ. Some consider 3 days inactive as churned, some 7 days, some 30 days; some look only at login behavior, some also at payment. Inconsistent definitions make the churned-user count differ by 2 to 3x across definitions. Every churn-profile query means rebuilding the filter criteria from scratch, averaging 20 to 30 minutes and impossible to compare across queries.
What it does
When to use it
A game project needing to define the "churned user" standard so the whole team uses one definition
An operations team needing to quickly build a churned-user cohort for win-back pushes and later analysis
Needing to describe churned-user profile characteristics such as channel, version, device type, and payment tier
A tool product needing to define the no-longer-using threshold
An e-commerce project needing to identify the no-longer-repeat-buying user group
In the field
FAQ
How many days should the churn definition use?
It depends on product type: 3 days for light casual/match-3, 5 days for mid-core RPG/card, 7 days for heavy SLG/MMO, and for tools it follows the usage frequency of key features.
Can this Skill tell me why users churn?
No. It only identifies and profiles. Churn-cause attribution needs other Skills, such as Level Churn Diagnosis or Payment Attribution.
How do I use the cohort once built?
You can reference it directly in AE models like event analysis and retention analysis, and use it to target audiences for win-back pushes.
Related Skills
Equip your Agent with Churned User Identification and Profiling
Book a demo and see how it works in your own business.
