A much overlooked issue in Customer Data Integration projects is “Persistent Identification”.
Persons and companies are very often identified using their address data. But, what do you do if a person has moved from address A to address B. One, thing you really don’t want is that the person is added to database as a new person (INSERT). From that moment a duplicate person or company resides in your system. This should be prevented, by creating searching indexes which include the current and the previous address of the persons and companies in your database.
A common way to solve this is to store all addresses which the person is known (historical addresses). For each address the start-date (FROM …) and end-date (TILL …) should be stored. Especially techniques which are used in Data Warehousing, like dimensional modelling, come into place.
Of course, you also need other identifying attributes to recognize the moved person. Focus on other customer touchpoints like the e-mail address of persons. Nowadays an e-mail is less subject to change then the physical address where somebody lives. A social security number or banking account number are even more precise, but unfortunately in many cases not available.
And, last but not least, if possible make uses of external sources. For instance in the Netherlands the movers index is maintained by TPG (TPG Verhuisberichten). In other countries similar services are available.