TIBCO Software
363 Case Studies
A TIBCO Software Case Study
AT&T tested a new direct long-distance billing strategy to weigh its risk to customer retention. In a controlled experiment customers either continued receiving long-distance charges on their local phone bill (control) or got a separate AT&T bill (treatment); survival analysis (Weibull regression and Kaplan–Meier curves) was used to compare switching risk across high- and low-usage segments.
The analysis showed the treatment group had similar early switching risk but, particularly in the low-usage segment, a significantly lower long-term cumulative hazard—suggesting direct billing can improve retention. A density plot of reversion times revealed two spikes within the first ~50 days (the first two bills), so AT&T should concentrate marketing and education efforts in that initial period to minimize early churn.