Combining systems that disagree.
WHAT THE PROBLEMS ARE
The same entity identified differently in each system Different definitions of the same concept Conflicting values Different update frequencies Different levels of quality
WHAT ENTITY RESOLUTION IS
Determining that records in different systems refer to the same thing.
WHAT MAKES IT HARD
No shared identifier Names spelled differently Values entered inconsistently
WHAT APPROACHES EXIST
Exact matching on a shared key, where one exists Deterministic rules on combinations of attributes Probabilistic matching, scoring similarity Manual review of uncertain matches
WHAT TO BUILD
A mapping table recording which records correspond, and how that was determined.
WHY A TABLE
So decisions are visible, correctable, and stable.
WHAT TO DO WHEN SOURCES CONFLICT
Establish which is authoritative, per field.
WHY PER FIELD
One system may be authoritative for contact details and another for financial status.
WHAT TO DOCUMENT
That decision, explicitly.
WHAT TO RETAIN
The source values, alongside the resolved one.
WHY
So a wrong resolution can be traced and corrected.