Use survival analysis to estimate churn-free tenure when some customers are still active. Learn how hazards, censoring, and model choices change your retention insights.
Which method estimates retention while properly handling right-censored customers?
Principal component analysis
ARIMA
K-means clustering
Kaplan–Meier estimator
In a churn context, the hazard rate at time t represents ______.
average tenure across the cohort
the instantaneous risk of churn at time t given survival until t
the lifetime probability of ever churning
instantaneous revenue per user
The key assumption of a Cox proportional hazards model is that ______.
there is no censoring in the data
all covariates are time-invariant by definition
covariate effects multiply the hazard and remain proportional over time
survival times are normally distributed
To incorporate monthly activity that changes over time, you should use ______.
simple linear regression on tenure
a Cox model with time-varying covariates or a piecewise approach
naive Bayes on churn labels
Kaplan–Meier with static strata only
The median tenure from a survival curve is defined as ______.
the time when the estimated survival falls to 0.5
the mean of uncensored durations only
the mode of the hazard function
the first time hazard equals 0.5
Right-censoring occurs when ______.
event time is known to lie within an interval
a subject has not churned by the end of observation
entry time is unknown (left-censoring)
observations are truncated before entry
Restricted mean survival time (RMST) is best used to ______.
remove the need for censoring adjustments
summarize expected tenure up to a fixed horizon
test the proportional hazards assumption directly
estimate instantaneous churn risk
If churn can happen via distinct, exclusive causes (e.g., voluntary vs. involuntary), an appropriate framework is ______.
standard Cox PH without cause indicators
binary logistic regression only
KM ignoring cause information
a competing risks model (e.g., cause-specific hazards or Fine–Gray)
Accelerated failure time (AFT) models are typically interpreted via ______.
time ratios describing multiplicative effects on survival time
odds ratios of churn at any time
differences in mean tenure only
p-values of Schoenfeld residuals
Schoenfeld residual checks are commonly used to ______.
decide between right- and left-censoring
estimate the baseline hazard function
assess violations of the proportional hazards assumption
compute RMST directly
Starter
You know the basics of retention curves and censoring.
Solid
Good grasp of hazards and when to use different survival models.
Expert!
You can choose, diagnose, and explain survival models to stakeholders.
