Party Master Data – names and addresses – is the most common entity in all the worlds databases appearing as tables with customers, prospects, vendors, members, contacts or whatever description suitable for the task at hand.
I have often found it useful to typify the names and addresses to be matched. I have used this extended ABC model for that:
A is a pure ADDRESS. Here the name is either empty, is not important at all or you are not (temporary) able to typify it more precise with the following types.
B is BUSINESS. This is a name and address of a business entity or in fact any organisation at all being a legal entity. Be aware that these may be further organised in hierarchies, e.g. local branches, domestic headquarters and world wide enterprises.
C is CONSUMER (or CITIZEN). This is a private person name on the private residential address.
D is DEPARTMENT. This is a non legal division of a BUSINESS often introduced as a billing or delivery address.
E is EMPLOYEE. This is a named person belonging to a BUSINESS like in the classic CRM data model a contact under an account.
F is FUNCTION. Like EMPLOYEE but with no personal name – here only a job function or decision level description is available within a BUSINESS.
G is GROUP. Here you have 2 or more names in the same string, e.g. “Mary & John Smith”. If not recognized as a BUSINESS these could be split into 2 or more CONSUMERS – or typified as a HOUSEHOLD.
H is HOUSEHOLD. This is in fact a fuzzy entity but never the less an important entity to many operations as being 1, 2 or more CONSUMERS at same residential address with a shared budget.
I is INVALID. This is all kind of dirt, fraud, test, comment and other illegible names and addresses that are not within the purpose of the domain. You are always amazed how many records that end up here.
A few points on the purpose on these types and how they may be established:
- These types may often support the transition from one data model to another.
- These types may serve as an important criteria in deduplication as you may avoid merging different types and within some of the types.
- Establishing these types may include intensive use of external reference data as business directories and other directories on addresses, consumers/citizens, names and more – and the costs as well as general availability, coverage, depth, actuality and other metrics differs a lot from country to country.