Overview
Dynamics 365 Customer Insights enables users to unify data from multiple data sources and convert it into a single master dataset providing a unified view of customers. By leveraging its AI-based predictive matching capability, users can match, map and merge any kind of data, define multiple configurable match rules, and even set a threshold of matching precision to have granular control. In this article, I’ll go through the steps involved in data unification.
Pre-requisite
Before unification, the data ingestion step should be completed, and the data to unify should be available in the system. The system is capable of getting and unifying data from multiple data sources.
Map
To begin data unification, you need to select the entities you want to unify.
Go to Data->Unify->New entity
You can select multiple entities from multiple data sources available in the system.
For example, I am adding Customer, InvoicesCRM, and InvoicesWH. These entities correspond to each other with the customer’s last name and product number, respectively.
Once entities are added, one can review/change the type of attribute inferred by the system, add/remove attributes, and select the primary key.
Note that the system is smart enough to infer several different types of attributes like city, country, email, and several others.
Match
Once the Map process is complete, you can then start the match process. In this step, you will define rules to specify how selected entities are related to each other and which fields can be used for matching.
Match Order
First, you need to specify match order, go to Data->Unify->Match->Set order
The screenshot below shows the match order and how the system will match according to the defined order.
Once the match order has been set, you can then define rules to set conditions based on which attributes will be matched in participating entities.
Match Rules
On the Match panel, click on the Add rule button to define the rule and add conditions. In the below screenshot, the condition is defined to match two entities based upon the Last Name. You can add more conditions by clicking on “Add Condition.” The Precision level value can be set from the slider. If you select the custom precision level option, you can define the exact matching threshold.
You can define multiple matching rules, and each rule can contain multiple conditions.
Validate and Review
After defining rules, you can click on “Run” to start the match process. Upon successful completion, it’ll then create a unified entity with the name “ConflationMatchPairs : CustomerInsights” on the Entities page.
You can then review the data to evaluate the match process and fine-tune the rules’ precision threshold and conditions if required. You can also edit conditions of existing rules or add more conditions and rules. The match preview of a specific rule can be viewed from the button highlighted in the screenshot below.
The match panel will provide you with necessary information about the rules, such as their order, how many records are matched according to defined rules, and the total unique customers identified by the system.
The system also provides custom match capability based upon the user-defined template.
Merge
In this last phase of data unification, conflicting data can be reconciled, and you can review and decide the attributes of the customer profile (a resultant entity of merge). All the attributes of participating entities are available, and the user can decide which attributes to exclude from the customer profile.
Once you run the merge process successfully, it will generate unique customer profiles based upon selected attributes. A corresponding entity, “Customers: CustomerInsights” will be available in the “Entities” page.
Conclusion
With AI-based data matching capabilities of Dynamics 365 Customer Insights, you can map, match, and unify any kind of data, which can be very time-consuming if processed manually. It provides a simple, intuitive yet powerful interface to configure and perform all steps involved in data unification with granular control.
If you would like to know more about data unification, or have any questions on the topic, feel free to leave a comment below.
Happy unification!
