Whether the multinational system is to operate in one country and accept data from multiple countries, or whether the system is to be deployed in multiple countries, the way that names and addresses are captured and stored is crucial to the reliability of future matching on that data.

Names and addresses from different countries have different structures, follow different rules and differ in average quality. In Canada, it may be difficult to get a letter delivered without a post code; in Hong Kong, almost no one uses the post code.

A data model which assumes that the data for each country can be mapped into a detailed universal name and address format look nice on paper in the specifications, but will be costly to implement and generally unsuccessful in practice. The universal format for a name is a single field holding all name parts. Simply make sure the field is big enough.

The universal format for an address is multiple lines, as written on an envelope. Simply make sure the field is big enough.

If the success of matching name and address data in your multinational system is important, do not trust match keys or matching logic which rely on the data being parsed, cleaned or formatted.

Use simple large single fields for name data, and a box of multiple lines as is used on an envelope for addresses. A search and matching system, which succeeds with the full unformatted name and unparsed address lines, will be easier to implement, more flexible and ultimately give more reliable results.