The following are notes that I took live while Andrew Grenville (Chief Research Officer, Vision Critical) and Kevin Dang (Senior Research Manager, Vision Critical) presented at MRIA Netgain 7.0. The notes were originally posted on my Tumblr account (paullongsblog.tumblr.com), but I have moved them here and revised them slightly to remove typos.
Data Mining and Predictive Modeling to Drive Panel Management Strategies
Analytical Panel Management — Key Factors
- Data set-up
- Business Challenge
- Methodology and Analysis
- Evaluation and Performance
- Application
Analytic Maturity Curve
Steps (starting from bottom)
- Static reports
- Drill down
- Ad hoc reports
- Forecasting
- Predictive analysis
- Optimize
Developing Analytic Capability
- Demographics
- Activity
- Incentives
- Value
Want to find out the overlap between organizational and human vectors
- Getting buy in from organization
- Clean data
- Get help from IT
- Have the skill set
Managing Attrition Using Analytics
ARF Churn Propensity Model
- Panelists are purged after 6 months of inactivity, 1 month lag time, and then one month to adjust the strategy so a total of an 8 month timeframe.
After building model
- 76% accurate model, idea was to target the folks scoring 7-10 on likelihood to churn.
Challenge: Create the least complex model
Can target high risk churn group, test the retention strategy, test effectiveness and refine, and then start again.
Have not finalized how to do this, but there are monetary and non-monetary elements, newsletters.