Click here for a printer friendly version
Retention
Predictive Data Modeling Helps Identify Likely-To-Leave Customers — Telecommunications
The Challenge
Our telecommunications client determined that a significant number of its business service contracts had the potential to expire between November 2003 and March 2004.
- For the November/December timeframe, the erosion rate was projected at 20%.
- Thereafter, the erosion rate was projected between 18% and 24%.
The client needed to Implement an immediate, proactive outbound retention marketing initiative to prevent a “waterfall” attrition of services customers. An acceptable goal would be to reduce the attrition rate for the pilot November/December time period by 2% to 4%.
The Breakthrough
SIGMA recommended a three-phase strategic knowledge mining, predictive modeling and analytical service process that would:
- Provide a predictive data profile for likely-to-attrite business contracts;
- Prioritize the 65,000+ customers that expired in November/December according to that predictive data profile.
Phase I – Q4/2003
Use predictive data modeling to develop customer profiles for most likely-to-attrite accounts.
- Develop observation scoring for list prioritization as the basis for proactive retention-based telemarketing to customers.
Phase II – Ongoing Retention Strategy – Q1/2004
- Conduct a Phase I results analysis and incorporate those learnings into the ongoing strategy.
- Identify attrition triggers and use predictive modeling for ongoing targeting of most likely attritors.
Phase III – Cross-sell/Upsell and “Code Red” Automation 2004
- Conduct a Phase II results analysis and incorporate those learnings into the ongoing strategy.
- Develop “next best product/service buy” predictive data models to cross-sell/up-sell current clients.
- Develop a Web-based delivery infrastructure for automated retention program lead management.
The Win
- Of those customers who did cancel their accounts during the pilot timeframe, the SIGMA profiling accurately identified – within the top two deciles – 82% of those customers. This provided validation of the predictive strength of the profile.
- Through a proactive, outbound communication program to the at-risk accounts, the 20% projected attrition rate was reduced to just 4%.
- A significant percentage of the not-likely-to-renew accounts not only stayed, but also extended the terms of their current contracts beyond the one-year standard.