
Marketing
Reducing Customer Churn
Build a high-performing customer churn model to improve customer loyalty, realize sustainable revenue growth, and optimize business processes.
Challenge
- Historical customer data is dispersed in multiple system environments and cloud providers across jurisdictions.
- Data centralization commingles user data, requiring operations to anonymize and duplicate data which reduces signals and increases storage costs.
- Data science teams struggled to quickly normalize data to report on critical business objectives, often resulting in high turnover rates for data science talent.
Solution
- Devron’s forward-deployed engineers help data science teams deploy and train satellite models on data at each of the provider’s 23 regional offices.
- Insights from satellite models are aggregated and the updated models sent back out so models at every data store can learn from each other.
- Budgets reallocated to decentralized (federated) data science initiatives following trends in cost reductions and performance of highly accurate ML models to reduce customer churn rate with machine learning.
Outcomes
↑
Increased Customer Retention
↑
Improved Customer Loyalty
↑
Greater Model Accuracy
↓
Reduced Data Centralization Time & Costs
↓
↓