Ques:
The business keeps changing metric definitions and logic. How do you structure your models in SAP Datasphere to support flexibility and adaptability?
Ans:
When the business frequently changes the metric definitions and the calculation logics, we can follow the approach that has central governance, low maintenance, and flexibility.
Here’s how we can structure the model in SAP Datasphere:
– Define all the metrics in the semantic layer of the business builder, and create and manage all the key KPI data there.
– This will help with business logic, which can be abstracted from physical data models, making it easier to modify.
– Now modify your data architecture and structure your models into layers like raw data layer, business logic layer, and semantic layer. The above separation will make it easy to update the logics without modifying the ingestion or base data models.
– We can reuse the views and calculations for creating intermediate reusable views, and these views can be referenced by multiple KPIs for reducing redundancy and easing the updates.
– Avoid hardcoding in the data flows to keep all the transformations and business logics out of the ETL logic and do the calculations in the views or semantic models instead.
– Using parameterization, wherever possible, to define flexible definitions.
– Implement versioning for metrics and create versioned KPIs or model variants if we want the business to compare the old and the new logic.
– Document all the metric definitions and use the business glossary in the business builder.
– Ensure all the changes to the logic are well communicated and auditable.
This way, we can deal with the business that frequently changes the metric definitions and the calculation logics.
For More Such QnA on Datasphere for Quick Revision, you can check out: https://topmate.io/vartika_gupta11/1639897