Customer, Supplier, and Location MDM Enhancements

Summary

The following enhancements to Customer, Supplier, and Location MDM functionality have been made as part of the 2024.4 update. These changes are outlined below and described in the Details section that follows:

  • Icons for all common object types within the Customer MDM data model have been added in the 'Select Icon' dialog.

  • A new Machine Learning Matcher version 2.0 for person name and address matching is released together with a new Address Normalizer v2, significantly improving the accuracy of matching.

  • The Address Component Model no longer contains mandatory fields, allowing users to input addresses with varying levels of detail without being restricted by field requirements.

  • The Machine Learning Matcher now offers the capability of processing sets of input data allowing, for example, multiple addresses to be evaluated and returning the highest score for more accurate matching results.

Details

Object type icons added

In the workbench, when configuring an icon for an object type, the 'Products / Entities' category listed in the icon library will now be separated into two different categories of 'Product' and 'Entities'. In addition to the icon library organizational change, additional icons have been added in the 'Entities' category to account for object types commonly included in the CMDM data models. These changes will improve the look and usability of the icons displayed in both the Web UI and the Instrument Web UI.

New ML Matcher version 2.0 for address matching and Address Normalizer v2

With the 2024.4 update, the Machine Learning Matcher version 2.0 has been released, featuring matching capabilities for both person names and addresses. To support this matcher, a 'StandardizedAddress' object type is available in JavaScript SDK, and a new ‘Address Normalizer v2’ has also been released. The previous Address Normalizer has been renamed to ‘Address Normalizer v1 (superseded).'

Whereas the initial 1.0 version of the Machine Learning Matcher focused on person name matching, version 2.0 expands these capabilities to include address matching. Leveraging Stibo Systems' machine learning functionality, the matcher handles challenges such as inconsistent address formats and incomplete data, resulting in more accurate address matching.

The new Address Normalizer v2 supports both input values and standardized values. Users of SaaS v2 systems are encouraged to move away from using the legacy person name and address matchers and transition to the Machine Learning Matcher and 'Address Normalizer v2' for enhanced functionality and ongoing support. The previous ‘Address Normalizer v1 (superseded)’ is still available as the only option for on-premises deployments.

To learn more about this update, click the video below. If the video does not play as expected, it can also be found in the Customer / Partner Communities, and may also be accessible within the Stibo Systems Service Portal.

For more information, refer to the topics Matcher: Machine Learning Matcher and Data Element: Address Normalizer v2 in the Matching, Linking, and Merging documentation.

Removal of mandatory field requirements from the Address Component Model

The Address Component Model no longer imposes mandatory field requirements. Previously, the model included specific fields that needed to be completed to meet the requirements of address validation systems such as Loqate. With this update, the Address Component Model now functions effectively without needing specific fields to be filled in, making the system more adaptable to different data sources and address formats.

For more information, refer to the topic Address Component Model and its subtopics in the Data Integration documentation.

Machine Learning Matcher now supports sets of input data

The Machine Learning Matcher now supports processing sets of input data, such as addresses. When comparing e.g., one record with three addresses to another record with two addresses, all six combinations are evaluated, and the highest score is returned as the result. To prevent service failures, the system allows a maximum of 20 elements of input data. If this limit is exceeded, a warning will be logged in the STEP log.

For more information, refer to the topic Matcher: Machine Learning Matcher in the Matching, Linking, and Merging documentation.

Stibo Systems MDM Connector for Salesforce

The MDM Connector is a Salesforce app that supports functionalities within Salesforce, such as bi-directional flow and synchronization of data, and ‘Merge’ and ‘Search before Create’ scenarios.

The app contains a managed package maintained by Stibo Systems, and a reference implementation supporting Accounts and Contacts that can be changed and extended by the customer.

The new MDM Connector app accelerates time to value and improves implementation experiences through a rapid and certified integration approach with Salesforce. A certified Salesforce app, coupled with a best practices reference implementation, empowers fast and efficient implementation.

The MDM Connector app works with any supported version of STEP. At present, the app is only available under an early adopter license. As an early adopter, you will work closely with Stibo Systems’ experts and the product management team in leveraging the MDM Connector for Salesforce and shaping the direction of the initiative moving forward.

To learn more about this update, click the video below. If the video does not play as expected, it can also be found in the Customer / Partner Communities, and may also be accessible within the Stibo Systems Service Portal.

Express your interest and learn more about the early adopter release of the Salesforce connector by emailing sfdcconnector@stibosystems.com.