Constructing a Pattern - trillium_quality - 2.7

Trillium Quality for Dynamics Guide

Product type
Software
Portfolio
Verify
Product family
Trillium
Product
Trillium > Trillium Quality
Version
2.7
Language
English
Product name
Trillium Quality
Title
Trillium Quality for Dynamics Guide
Topic type
Administration
Overview
How Do I
Configuration
Reference
Installation
First publish date
2008

After individual field comparisons are completed, each comparison is assigned a match score (0 through 100) and a grade (A through E) indicating how closely the fields match. The grades are then strung together to create a pattern. This pattern is compared to the pattern matching rules which will indicate if the pattern results in a match of the two records.

The pattern matching process is the same for accounts, contacts, and leads but different fields are used in the comparison. The following example uses contact records to demonstrate the pattern matching process.

Record #1

Record #2

Mr Bob Miller

123 Maine St

Billerica, MA 01821

Mr Robert Z. Miller

123 Main St

Billerica, MA 01821

The field comparisons, match scores and grades for these two records would be:

Field Compared

Record #1

Record #2

Match Score = Grade

Street Name

Maine

Main

95 = A

Last Name

Miller

Miller

100 = A

House Number

123

123

100 = A

Apartment Number

Blank

Blank

98 = C

Box Number

Blank

Blank

100 = A

Postal Code

01821

01821

100 = A

City

Billerica

Billerica

100 = A

House Number

To Apt Number

123 to blank

123 to blank

0 = D

Name Prefix

Mr

Mr

100 = A

First Name

Bob

Robert

80 = D

Middle Name

Blank

Z

98 = A

Generation

Blank

Blank

100 = A

Gender

M

M

100 = A

Standardized First Name

Robert

Robert

100 = A

The assignment of the score to a grade allows for a variance, or range, of how similar the field's values are to each other. This provides the flexibility to be able to weight the results of the comparison routines at the field level.

The next step in the matching process is to combine all of the calculated letter grades to produce a pattern. For example, the pattern produced from the example above is:

AAACAAADADAAAA

This pattern is then compared against the table of patterns, which will determine if the pattern indicates that the records are similar enough to be considered a match. For example, the pattern above could be determined to be a match because the standardized names of Bob and Robert living at similar addresses would be similar enough to be considered a match. The two records are flagged as a match and a Pattern ID (a number corresponding to a combination of matching attributes and grades) will be assigned to them.

This comparison process allows weighting of the entire record, which provides the flexibility of matching records based on differing levels of field similarities. It also allows for full auditing of why the records were matched.