How
do you find your customers in your data bases when you have only
their contact information?
About
Mematch
Imagine identifying your customers third
party data by name, address, telephone # and other contact info
against your administrative customer databases.
How would you efficiently match slight
variations on your customers names and address against your master
administrative database standard name and address?
How can you set up an efficient process
that uses different matching criteria to cross match and identify
your customers?
Business Case
- Client has a file with customer contact info that needs to be
cross referenced against another file having a field of interest
such as “Customer ID#”.
- Since contact info is not kept in a standard format and may
have slight variations in both files, a simple match by “Name”,
“Address”, “City”, “Zip Code”
and “State” or other information may produce only
a few matches.
- Manual variations of “Name”, “Address”,
“City”, “Zip Code” and “State”
or other information will increase # of matches but are time consuming
and error prompt.
The MeMatch Solution
- The algorithm matches variations of contact info in both files
to produce a file with the corresponding customer ID.
- The algorithm allows for the use of the sounds like criteria
The algorithm criteria is table driven. It assumes that the criteria
Rank, crit field, in the criteria table indicates the accuracy
and strictness of a given criteria when compared to others. For
example It assumes that crit #1 is the most accurate and strict,
# 2 the second most accurate and strict.
- The algorithm goes in sequential order and loops through "n"
number of matching criteria in the criteria table. A criteria
table outlines the rule for matching records. For example a criteria
may require exact matches on name, address, city, state and zip
code. Another criteria may required matches on the first 7 characters
of name, and full matches on address, city, state and zip.
Business Case
- Client has a file with customer contact info that needs to be
cross referenced against another file having a field of interest
such as “Customer ID#”.
- Since contact info is not kept in a standard format and may
have slight variations in both files, a simple match by “Name”,
“Address”, “City”, “Zip Code”
and “State” or other information may produce only
a few matches.
- Manual variations of “Name”, “Address”,
“City”, “Zip Code” and “State”
or other information will increase # of matches but are time consuming
and error prompt.
The MeMatch Solution
- The algorithm matches variations of contact info in both files
to produce a file with the corresponding customer ID.
- The algorithm allows for the use of the sounds like criteria
The algorithm criteria is table driven. It assumes that the criteria
Rank, crit field, in the criteria table indicates the accuracy
and strictness of a given criteria when compared to others. For
example It assumes that crit #1 is the most accurate and strict,
# 2 the second most accurate and strict.
- The algorithm goes in sequential order and loops through "n"
number of matching criteria in the criteria table. A criteria
table outlines the rule for matching records. For example a criteria
may require exact matches on name, address, city, state and zip
code. Another criteria may required matches on the first 7 characters
of name, and full matches on address, city, state and zip.
|
