This article explains the predefined and custom duplicate detection logic in Sumac, which helps identify duplicate contacts based on configurable sensitivity levels.

Predefined Duplicate Detection Logic

Sumac offers three predefined levels of Duplicate Detection Sensitivity: Lenient, Balanced (default), and Strict.

1. Lenient

   1. Partial Match Count: 3

   2. First Name Logic:

      1. All letters match: Partial Match

      2. First letter match: Partial Match with 4 letters

      3. Letter difference match: Partial Match with 1 letter

   3. Last Name Logic:

      1. All letters match: Partial Match

      2. First letter match: Partial Match with 4 letters

      3. Letter difference match: Partial Match with 1 letter

   4. City Name Logic:

      1. All letters match: Exclude Match

      2. First letter match: Exclude Match

      3. Letter difference match: Exclude Match

   5. Phone Logic:

      1. All letters match: Partial Match

      2. First letter match: Partial Match with 5 letters

      3. Letter difference match: Partial Match with 1 letter

   6. Email Logic:

      1. All letters match: Full match

      2. First letter match: Partial Match with 5 letters

      3. Letter difference match: Partial Match with 1 letter

   7. Include match between Organization and Individual: Don't include
      

2. Balanced (Default)

   1. Partial Match Count: 3

   2. First Name Logic:

      1. All letters match: Partial Match

      2. First letter match: Partial Match with 5 letters

      3. Letter difference match: Exclude match

   3. Last Name Logic:

      1. All letters match: Partial Match

      2. First letter match: Partial Match with 5 letters

      3. Letter difference match: Exclude match

   4. City Name Logic:

      1. All letters match: Exclude Match

      2. First letter match: Exclude Match

      3. Letter difference match: Exclude Match

   5. Phone Logic:

      1. All letters match: Partial Match

      2. First letter match: Partial Match with 5 letters

      3. Letter difference match: Exclude match

   6. Email Logic:

      1. All letters match: Partial Match

      2. First letter match: Partial Match with 5 letters

      3. Letter difference match: Exclude match

   7. Include match between Organization and Individual: Don't include
      

3. Strict:

   1. Partial Match Count: 4

   2. First Name Logic:

      1. All letters match: Partial Match

      2. First letter match: Exclude match

      3. Letter difference match: Exclude match

   3. Last Name Logic:

      1. All letters match: Partial Match

      2. First letter match: Exclude match

      3. Letter difference match: Exclude match

   4. City Name Logic:

      1. All letters match: Exclude Match

      2. First letter match: Exclude Match

      3. Letter difference match: Exclude Match

   5. Phone Logic:

      1. All letters match: Partial Match

      2. First letter match: Exclude match

      3. Letter difference match: Exclude match

   6. Email Logic:

      1. All letters match: Partial Match

      2. First letter match: Exclude match

      3. Letter difference match: Exclude match

   7. Include match between Organization and Individual: Don't include
      

Custom Duplicate Detection Logic

How Duplicate Detection Logic works:

1. Partial Match Count - is a minimum number of Partial matches for marking Contact as duplicate.

2. Full match is a type of match when Contact is marked as duplicate.

3. Partial match is a type of match when this match is counted for Partial Match Count check.

4. Exclude match is a type of match which is excluded from the logic

5. Letters count for First letter match is a number of first letter to be checked for an appropriate field.

6. Letters count for Letter difference match is a maximum distance between field values calculated using Levenshtein distance algorithm (Levenshtein distance ).

7. Include match between Organization and Individual is a flag whether or not compare Contacts which have Organization/Household gender with Contacts which have individual gender. If “Don't include” is set Organization/Household contacts will be only compared with Organization/Household contacts, and Individual contacts will be only compared with Individual contacts.

8. If Contacts have at least one of Full matches or Partial Matches Count greater or equal Partial Match Count they are marked as duplicated.