Match Rules in Data Cloud

One of the core value proposition of Data Cloud is data harmonization-- a fancy term for consolidating multiple profiles from disparate data sources into a single 'Unified Profile'. Why is this important? Clean data is foundational to effective and accurate AI endeavors. The training data for Machine Learning will be riddled with inaccurate samples, duplicate rows,etc. Simply put..bad data begets bad AI.

User Story: Company XYZ has an SFDC instance with 3 customers in Contacts -- William Hall  (email:[email protected]),  William Hall (email:  [email protected]) and Will Hall (email:[email protected]). Through proper configuration of Match and Reconciliation rules, Data Cloud (formerly known as Customer Data Platform or CDP) will be able to consolidate the 3 Mr. Halls into the cleaner version consisting of 2 William Halls -- the William Hall with the gmail address and the William Hall with the aol.com address.

The match rule that will trigger the above harmonization is 'exact last name+fuzzy first name+exact email' . Additional rules can be layered to this such as 'exact frequent flyer miles' or 'exact driver license number',etc. to further 'unify' profiles-- more rules added means more consolidation (ie increase consolidation rate).  The snips below illustrate this... 1st set of snips has 4 match rules -- Driver License (DL) OR car club OR exact last name+email OR frequent flyer. This set unified 200 profiles into 196 . The 2nd set added a 5th match rule 'Motor Club" which means if a person has the exact motor club Id in 1 or more objects involved in the harmonization, they are to be unified into 1 profile.