In your scenarios, choose a matching method that best suits the fields you are comparing.
Matching Methods are algorithms used by Plauti Deduplicate to analyze field values. Based on this analysis it decides to what extent two records match: the matching score.
Record fields can have different value types (e.g. numeric or text, but also plain text or email addresses, etc.) In order to catch duplicate records it is important to apply a matching method that can properly analyze a field's value type. Plauti Deduplicate's matching methods are designed to do just that.
Furthermore, with a matching method you also decide to apply either an exact or fuzzy logic to evaluate field values.
-
Exact methods look for values that are almost exactly the same, sometimes with differences in casing or spaces allowed depending on the method.
With exact matching, the matching score is either 100% or 0%: two values match or they don't. -
Fuzzy methods allow for slightly less exact matching, so that spelling mistakes and conventions such as American vs British English can be taken into account.
Fuzzy matching uses a comparison system called Soundex to find similar values, and attaches a graded matching score based on the level of similarity.
Note that for Fuzzy methods, it is essential that the Search Index has been configured.
In future versions of Plauti Deduplicate for Dynamics 365, more matching methods will be made available.
Prerequisites
- For Fuzzy methods the Search Index needs to be configured.
Applying Matching Methods
Matching Methods are applied in Scenarios. You can select a different matching method for each scenario field, suitable for that field type.
Create the Search Index
In order to use Fuzzy matching methods, the Search Index needs to be created for each Object where you want to use them. After setting up or editing a scenario, create or re-create the Search Index for that Object before starting the first duplicate search.
The Matching Methods explained
| Matching Method | Specifics |
|---|---|
| Company Name |
Company Name is a fuzzy matching method that will flag values as duplicate when a word in one string is found in the other string, but allows for spelling mistakes or different formats.
Examples
|
| Domain |
The Domain matching method is an exact matching method that checks if two URLs or email addresses have the same domain name.
Example 1 http://plauti.com all score 100% when comparing these values with each other. Example 2 jane.fielding@plauti.com all score 100% when comparing these values with each other. |
| Equal |
When using 'Equal' as a matching method, only exact matches will generate a score.
Example 1 "Sten" scores 0% with "Stan" Example 2 "jane.fielding@plauti.com" scores 100% with "Jane.Fielding@plauti.com""jane.fielding@plauti.com" scores 0% with "janefielding@plauti.com" |
| Equal Random Order |
When applying the 'Equal (Random Order)' matching method, two values will score 100% if they contain the same words, but potentially in a different order.
Example 1 "Ruth Bader Ginsburg" scores 0% with "Ruth Badar Ginsburg" |
| Fuzzy |
The Fuzzy matching method will flag values as duplicate when a word in one value is found in the other value, but allows for spelling mistakes or different formats.
Example 1 "Sten Ebenau" scores 100% with "Sten Ebennau" |
| Person Name |
The Person Name matching method uses fuzzy logic to compare person names. It combines comparison results with the Levenshtein distance, and returns the average score of these.
"Sten Ebenau" scores 90% with "Sten Ebennau" "Fielding" scores 74% with "Feilding" "Ginsburg" scores 88% with "Ginsberg" "Bader Ginsburg" scores 92% with "Bader Ginsberg" |
| Phone Number |
The Phone Number matching method is an exact method that can score 100% on phone numbers written in a different format or standardization.
Example 1 (415) 555-2761 all score 100% when comparing these values with each other. Example 2 (415) 555-2761 scores 0% with (415) 555-2762 |
Cannot select all matching methods
For some fields, such as 'Company Name' in a scenario for Contacts, you can only choose the Exact matching method.
These fields are "lookup fields" that do not contain a value themselves, but refer to a field on another entity. Therefore they can only be compared with the Exact matching method.
