Description
Compare groups defined by shared characteristics or behaviour to investigate differences in retention, activity and product use.
Compare groups of users defined by shared characteristics or behaviour to understand differences in product use. Use cohort analysis when an overall average hides differences between audiences or patterns of engagement. Define and save the cohorts, choose a comparison group and inspect retention, activity or user properties side by side. Look at how the groups overlap as well as how they differ. Use the comparison to identify which behaviours or audiences need further investigation.
Attributes
- Discipline
- Epidemiology & Population Health
- Loss type
- Active
- Stage
- Post-launch
- Data
- Quantitative
- Effort
- Hours
Steps
- Define the cohort grouping (e.g. sign-up week) and the outcome metric (e.g. weekly active, purchase).
- Build a cohort table with outcome rates at fixed intervals from each cohort's start.
- Plot retention curves and check where they flatten and whether newer cohorts differ.
- Segment cohorts by acquisition channel, plan or first-week behaviour.
- Interpret differences carefully, separating age, period and cohort effects.
- Test hypothesised drivers experimentally rather than relying on behavioural cohort correlations.
Sources
- Advancing into Analytics
- The age selection of mortality from tuberculosis in successive decades (Wade Hampton Frost)
Videos
Related talks
Also known as
- cohort analytics
- retention cohorts
- Retention cohort analysis
- Vintage analysis
Categories
- Analytics & Behaviour
- Epidemiology & Population Health