An overview of a clv project Jimmy hosang
Summary
CLV projects require multiple actions to deliver their goals. These are: 1. Engagement + Accessibility 2. Data + Governance 3. CLV Model Build + Prescriptive Analytics
Collaboration is key
Engagement with Subject matter experts across marketing, pricing, channel and customer operations is key
to creating accurate datasets
Clv as a hygiene level
Clv is a key performance indicator
Performance indicators – profit, loss
Operational indicators – sales,
retention
Building the data in a hierarchy allows ownership flows
Assess the team: we want specialists
sas + Sql+ r + python + BI + Segmentation + survival analysis + soft skills = our analysts
Assess the systems
Proc expand; Do we have sas 9.4?
Do we have enterprise miner? How fast are our servers?
How powerful are our pc’s? Run;
So back to the clv model…
Segmentation: defines what type of customers you have by the similarities between groups
Survival and proportional hazard: predicts likelihood to leave and probability to cancel
Build the model iteratively.
Start simple.
Generate outputs.
Build complexity.
Create feedback loops. A good clv model can take months. Keep engagement high with an iterative approach
And finally… prescriptive analytics
Sensitivity analysis: How does retention rate effect clv? What are the impacts of cross-sell/up-sell? How can we flex operational performance? What is our marketing effectiveness?