Product: ImportOmatic
Description: Instructions on how to auto-pick using the scoring option
Environment: All
Version: All


ImportOmatic can auto-pick a record based a scoring threshold setting.

In the left hand navigation pane, go to Duplicate Search and check the box to "Perform a duplicate search when a record cannot be matched by ID".

Next, in the left hand navigation pane, select "Duplicate Search" and then highlight  "Auto-pick". 

Check the box to "Auto-pick record from search results if a single record exactly matches" : Select Scor. This will activate the scoring meter toward the bottom of the settings view. Select the threshold you would like the scoring to recognize and adjust to suit the import you will be performing:


If only one match is found based on the duplicate search criteria and that record meets or exceeds the duplicate similarity score threshold setting, IOM will automatically select that record.

Duplicate similarity score

Each record returned in the duplicate search results is assigned a duplicate similarity score.  This score is a measure of how close of a match that record is based on an evaluation of several fields.   


When selecting the auto-pick score option use the slider to select a score value; if one and only one record is returned by the duplicate search results, auto-pick will automatically select that record if it meets or exceeds the value you have set.  


The duplicate similarity score evaluates various fields for how closely they match and weights the fields for how much of an impact each has on an overall confidence for matching.  For example, a record that has a blank first name and matching last name may score lower than a record with a matching first name variant (eg, Bob vs Robert) and matching last name, and it may score higher than a record with an obviously different first name and matching last name.  Likewise, an exact match when only first name and last name are mapped may score differently than a mismatch on first and last names with an exact match on address, phone, email, birthdate, gender, etc.


Records that are close but not an exact match may score highly but they will not score 100.  For example, matching the first name to the nickname will not score as highly as matching first name to first name.


Duplicate similarity scores are always calculated the same way – using exclusions will not alter the scores.  Exclusions may prevent a highly-scoring record from being auto-picked but the score itself is not reduced based on exclusions.
Please refer to the ImportOmatic User Guide for additional information.