Predictive Match Policy Type Definition

[This topic is pre-release documentation and is subject to change.]

A Predictive Match Policy is applied to certain hub data retrieval operations to obtain a list of potentially matching entities, of either Profiles or Interactions. For more information about predictive matching, including the REST operations that enable the creation and management of policies, see Predictive Matching Policy APIs (Hub). Customer Insights supports two types of predictive matching (also called conflation): Profile-to-Profile and Profile-to-Interaction. Therefore, the primary type is assumed to be "Profile"; the alternate type is either "Profile" or "Interaction", respectively.

Properties

Property names appended with asterisks* are required.

PredictiveMatchPolicy properties

Property JSON type Description
predictiveMatchPolicyName* string Match name
displayName map Localized display name
description map Localized description
primaryTypeName* string Entity type of the primary entity to be matched
alternateType* string ["Profile" | "Interaction"]
alternateTypeName* string
matchThreshold* integer Matching threshold, from 80 to 100 inclusive, in increments of 1
topMatchResultsCount* integer Number of matches above the threshold that we should be saved
alternateTypeResolutionId string Identifier of the list of property names that can be used as an identity for the predictive matched interactions
isRecurring boolean True if the match should be run on a periodic basis; false if match should only be run once.
lastMatchTime datetime Last time the match was run (readonly)
relationshipName string Name of the relationship created for storing profile-to-profile match results (readonly)
matchPropertyMappings* MatchPropertyMappings[ ] Primary-to-alternate property mappings, including hints. (see below)

MatchPropertyMappings properties

Property JSON type Description
primaryTypePropertyNames* string[ ]
alternateTypePropertyNames* string[ ]
matchHint string

The following properties can be updated after creation: matchThreshold, topMatchResultsCount and isRecurring.

JSON example

}
    "PredictiveMatchPolicyName": "socialactivity_customer_email_name",
    "PrimaryTypeName": "Customer",
    "AlternateType":"Interaction",
    "AlternateTypeName": "SocialActivity",
    "DisplayName": {
        "en-us": "Predictive match policy for Social-Activity interaction and customer (on name and address properties)"
    },
    "Description": {
        "en-us": "Predictive match policy for social-activity interaction and customer (on name and address properties)"
    },
    "MatchPropertyMappings": [
    {
        "MatchHint": "Email",
        "PrimaryTypePropertyNames": [
            "Email"
        ],
        "AlternateTypePropertyNames": [
            "UserName"
        ]
    },
    {
        "MatchHint": "Name",
        "PrimaryTypePropertyNames": [
            "Name"
        ],
        "AlternateTypePropertyNames": [
            "FirstName",
            "LastName"
        ]
    }
]