TUTORIAL: Aligning Big Data & Social Data to Generate Propensity Models
CEO & Co-Founder, LumenData
MDM systems are best at creating a unique, accurate, known view of the customer. The data in MDM systems is deterministic and great value is placed on accuracy, consistency, data lineage and so on. In a B2C setting, this includes household information, products owned, purchase history and so on. One of the new emerging (and additional) benefits of these systems is to derive insights into consumer propensities. In order to do so, master data - or sometimes what is known as “1st party data” - must be combined with ambient data, and then statistical algorithms applied to it. “Ambient data” includes source data (referring URL, device types, IP address, geo location) combined with “3rd party data” sources that can then offer great insight into consumer behaviors and propensities.
This tutorial discusses some of these techniques in the context of a customer case studies for whom these models and techniques were developed.
• Identifying & governing which Big Data/social/advertising data sources help comprise the “longitudinal” customer view
• Understanding how to map these data sources to 1st party MDM data
• Accommodating & addressing customer concerns of privacy & security when blending 1st & 3rd party data