Offer Management Strategy
Intelligent Offer Filtering, Ranking & Recommendation Engine for Customer Upgrade Journey
Intelligent Offer Filtering, Ranking & Recommendation Engine for Customer Upgrade Journey
Our existing upgrade process relies on manual segment management where staff manually select which offers are shown to each of ~70 customer segments. This approach is time-consuming, error-prone, and lacks personalisation.
We need an intelligent system that can automatically filter offers based on customer attributes, apply business rules, and rank recommendations in real-time across all channels. While we want to increase personalisation, simply creating more customer segments would make the manual management problem worse.
Note: The filtering examples above demonstrate the types of rules that can be applied. In theory, the system could apply multiple filters simultaneously based on customer attributes, credit limits, device preferences, and business rules. The actual filtering logic would become dynamic and customer-specific.
System identifies customers approaching end-of-contract using upgrade eligibility dates from existing SIM and handset base
All available offers are ingested and processed through the recommendation engine
Offers are filtered based on customer segments, technical eligibility, credit limits, and business rules
Filtered offers are ranked using business rules, customer preferences, and predictive models
Best offers are presented to customers through appropriate channels (app, CRM, outbound)
Apple customers see Apple devices only, Android customers see Android devices, network compatibility checks
Customers with low credit scores can only access offers up to certain spend limits, affordability checks
Price filtering based on churn likelihood and sleeper status, remove offers customers cannot afford
Different filtering rules for EARLY_UPGRADER, LIGHT_SLEEPER, SLEEPER, and HIBERNATOR segments
Customers who buy premium devices see premium device offers prioritised in their recommendations
Prioritise offers that increase data allowance, MRC spend, and overall value rather than decreases
Balance showing proven offers with testing new propositions using multi-armed bandit algorithms
Ranking algorithms learn from customer responses and adjust recommendations in real-time
Implement Pega Customer Decision Hub (CDH) or similar real-time decisioning platform integrated with Databricks. Alternative: Consider Hightouch Composable CDP as a modern, warehouse-native alternative.
Leverage existing Databricks infrastructure for ML model training, feature engineering, and customer 360 data. See Hightouch Databricks Integration for detailed approach.
Address cold start challenges for new offers using exploration algorithms and synthetic training data. Hightouch AI Decisioning provides built-in multi-armed bandit capabilities.
Connect decision engine to CRM, mobile app, and outbound campaign systems via APIs. Hightouch offers 250+ pre-built integrations for rapid deployment.
For a detailed comparison of Hightouch vs Pega CDH, feature mapping, architecture overview, and step-by-step implementation plan, see the comprehensive Hightouch Implementation Plan.
View Implementation Plan