About This Tool
I. Tool Overview
Behavior Addiction Level Analysis is a basic monitoring tool for continuous behavior, suited to games and mobile apps where you need to see whether users keep coming back.
It does not depend on a user's first day or prior history: any calendar day can be a cohort start. Users active on a given day form that day's cohort; the tool reports what share of that cohort is still active on the next day, two days later, …, day N, producing a cohort date × Day N retention matrix so you can quickly assess stickiness and addiction level.
Core Capabilities
- Cohort by calendar day: All unique users active on a date form that day's cohort; the date does not need to be each user's "first" day.
- Day N retention: For each cohort, (users still active on cohort day + N) ÷ cohort size, output as a percentage (2 decimal places).
- Triangular matrix: Rows = behavior dates (ascending), columns = Day 0, Day 1, …; one rate per cell for easy comparison across dates.
Use Cases
- Games: Logins, sessions, or payments aggregated by day—see what share of "active on day D" users are still active on later days; monitor retention and addiction.
- Mobile apps: Opens, core feature use, or check-ins by day—see continuous usage share for basic retention and stickiness monitoring.
- Other continuous behavior: Any events with user ID + behavior date can be aggregated by day and analyzed the same way.
Data Requirements
Upload Excel or CSV (UTF-8) with at least:
- User ID column: Unique identifier (e.g. UID, device ID, account ID).
- Behavior date column: Date of each event (e.g. 2026-03-01 or 2026/3/1).
- Behavior label column (optional): Type or name of the behavior; used for filtering or validation; calculation is by user + date (deduplicated).
Multiple events per user per day count as one; any day can be a cohort day—no need for "first day" or prior behavior.