Unified Customer Data Management
summary
Claims a warehouse-native customer data modeling pipeline that ingests behavioral and CRM data, resolves identity via hashed emails plus optional fingerprinting, builds interest profiles with embeddings (explicitly rejecting GPT for cost and determinism), clusters with K‑means and predicts 6‑month value using RFM‑based fits that perform quite well in holdout tests, and pushes audiences back into existing tools via reverse ETL and APIs (example: hashed payload to Meta) for real‑time personalization and campaign activation.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:49Banks and retailers need unified data
- 4:23Integrating customer data challenges
- 6:58Predictive marketing data hub solution
- 9:09Data sources and pipeline
- 10:59Identity resolution for users
- 16:49Customer interests and behavior modeling
- 20:24Customer value and lead scoring
- 24:43Activating data in marketing tools
- 26:07Syncing audiences with platforms
- 27:37Real-time personalization
- 28:12Summary
related talks
29mIdentity Resolution – The Brain of Customer Data Platform (CDP)
11mMaximizing personalization and AI models with quality customer data
06mUnifying Online and Offline Worlds
18mDemystifying Data Modeling
25mRevolutionizing Customer Experience Through AI & Data
18mBehavioral Data Creation for AI - Delivering Value and Insights
35mUsing AI to Fuel Digital Transformation & Growth
37mData Mastering at Scale with AI