Q: How to handle customer data integration without a master data management (MDM) system?
A: Start by identifying the business requirements and understanding the differences between active customers and prospects. Determine the common data points needed for integration and focus on answering key business questions. This approach helps define a roadmap for data integration and prioritize the most relevant data points.
Q: How to integrate customer data across systems for regulatory reporting without an MDM?
A: Instead of building a super customer table, create a lean data set that acts as a bridge between different data products. This data set should include identifiers and essential information for connecting customer data across systems. Alternatively, consider creating an aggregated data product that integrates customer and prospect data, serving as a single source of truth for specific use cases.
Q: How are data contracts implemented in practice? Are they executed at the source system level or through a gateway?
A: Data contract implementation involves two types of checks:
For batch processes, a stage-gate mechanism can be used. When a process finishes writing data, it triggers a signal. The contract evaluation engine then runs all the checks. If everything is fine, it signals downstream consumers to start. If not, it sends alerts for issue resolution.
Q: Are data contracts currently implemented only within the data warehouse, or are they also applied at the data source level?
A: Currently, data contracts are applied to the data lake and data warehouse. However, there are plans to expand their implementation to the operational plane in the future.