Using Contextual Identities in SAS Customer Intelligence 360
Introduction
SAS Customer Intelligence 360 (CI 360) allows marketers to define target segments based on varied selection criteria. For example, if they are part of a specific age group, household or hold a certain type of bank account. Contact and response events record interactions at the primary subject level (typically Customer). So, if another subject level (e.g. household or account level) was used to define the segment, then this information won’t automatically be available to downstream tasks.
Adding the context (that is, which information made this customer eligible for this offer) makes it possible to better understand how or why a customer was targeted. That information can then be used in campaign performance analysis and contact policy rules.
Consider the example of a bank where a customer could have one or more accounts. If the targeting criteria for an offer includes account level attributes (for example, account type), post-campaign analysis might include measuring performance by account type.
In CI 360, Contextual Identities add two standard attributes (type and identifier) to events.
- Context type provides the context level such as ‘account’, household’ or ‘subscription’.
- Context value contains the value of the identifier for the context type (for example, account ID).
For instance, a customer is being emailed because they hold a savings account where the account number is 12345 – the context type is “Account”, and the context value is 12345.
Adding Contextual Identities to a Segment
Contextual Identities can be added to segments created from On-Premises Direct data. To do this, a node must be marked as a segment. You then select the “Add Context to Segment” option and click on the “Select a Node” button. You will then see a selection window showing all the eligible nodes in the segment map that can be used to add context.
In this segment map below, I have created customer and household level criteria nodes in the same segment map. The customer node has customers with >50k income. The household node has households with expenditure <20k. I have marked the customer level as a segment map and added the context from the household node.
If you have customers in the customer node who are not in a household that meets the criteria of the household node, they will either be dropped from the customer node or have a blank context depending on whether or not you tick the “Remove Segment Members” checkbox.
A context node must meet the following requirements:
- The context node must be upstream of the segment node or in a disconnected set of nodes.
- The context node must be at a different subject level than the segment node.
- The relationship between the subject level for the segment node and the subject level for the context node must be one-to-one or many-to-one. If it is a one-to many or many-to-many relationship, you will be prompted to specify sorting criteria, with the first record being taken as the context to add.
For example, a customer could hold many accounts at a bank. Adding context to the segment of Customers could result in multiple context subjects (Accounts) matching a single subject (Customer) in the segment. If a customer has 2 accounts; account 1 has a balance of $100 and account 2 has a balance of $200. In this situation, the user would be prompted to use a sort to select the account to add the context for. If the sort order is based on balance descending, account 2 would be assigned as the context. Note that it may be necessary to refine the sort criteria and use multiple fields to ensure the correct record is selected.
Adding Contextual Identities to an Audience
Contextual Identities are automatically added to an Audience when the Audience is based on On-Premises Direct data and uses a Segment with context defined. In this scenario, Contextual Type and Contextual Value are automatically included in audience data. The contextual id attributes cannot be removed (but do not count towards the maximum 30 attributes that can be selected). The context type and context value will automatically be assigned in the audience definition as the subject and subject_id respectively.
For an API audience or audience based on an audience source, you select one of the data attributes to be the context column when creating the Audience definition. The context type is a entered as text and must be a meaningful value in relation to the context value field (e.g. Customer Segment).
There are some conditions that must be met:
- The context must be a different column than what was selected for the Identity column.
- This context settings cannot be changed after the audience definition is saved.
- An audience can have only one context.
Using the Contextual Identities in Tasks
These Contextual Identities can then be used in certain Tasks as follows:
Storing the Data
The Contextual Identity attributes “context type” and “context value” are added to several tables in the Unified Data Model (UDM) and Common Data Model (CDM) data from schema 16 onwards. Their exact column names are Context_type_nm and Context_val respectively.
For tables used by the tasks mentioned on the previous table, the context type and value will be populated as follows:
In addition, when tasks trigger contact and response events, the contextual information is collected in the CONTACT_HISTORY table and the RESPONSE_HISTORY table. You can use this data to determine why a contact or response was made.
Important: As the Context Values are stored in the UDM, do not select personally identifiable information (PII) as contextual Identities.
For the On-Premises Direct Segments, direct, the field will always be the subject_id field for the primary subject. It is therefore recommended to not to use a field like ‘Account Number’ as the subject id but rather to generate a surrogate key/identifier (like Account_id) when populating the marketing data mart.
For non-segment based audiences, there is no restriction on which field you designate as context value (other than it can’t be the same as the identity column). Care should therefore be taken when creating the audience to ensure that is not a PII value. To ensure data security, hashing functions could be used prior to sensitive fields being included as the audience context.
Further information
More information can be found on the SAS Customer Intelligence 360 User Guide available here.
Contextual Identities will be discussed in detail in a future update of the SAS Customer Intelligence 360 learning subscription. Thanks go to Carol Carlton and Maria Crossland for helping to compile this post.
Find more articles from SAS Global Enablement and Learning here.
