An enterprise’s business success depends on how efficiently it manages its crucial data named as Master Data. An Enterprise Data Warehouse is clearly one answer to this particular need, which is driven by some basic intentions:
- Need for historical data: The transactional systems would purge older data; master data would be replaced by newer definitions, thereby wiping out the previous versions.
- Need for increased data analysis performance: Transactional systems are not optimized for Data Analysis; complex analytical queries will impact day-to-day operations.
- Need for good and reliable data: Data from across different sources need to be cleansed, standardized and integrated for a good analysis. This clearly points at the need for “Single version of truth” concept, which is attainable by providing enterprise wide data consistency, accuracy and completeness. In fact this third point demarks an EDW from a normal data-mart based data warehouse.
However, as an enterprise grows both depth and breadth by Mergers and Acquisitions, thereby integrating newer sources with diverse standards, maintaining data consistency, quality and integrity become increasingly difficult task. Despite having stringent business rules and integration “dos and don’ts” enterprises land into multiple versions of its master data spread across LOBs.
Here comes the role of MDM that can be leveraged to achieve this enterprise wide “single version of truth”. Basically, MDM along with Data Governance and right Stewardship would help us define and enforce data policies, data quality across the board.
Ideally both approach for information management ultimately aims at the same goal: to realize an enterprise wide vision to company’s business data, however, there are some differences the way both are looked at and the way one could be leveraged by the other in order to successfully realize the power of enterprise data vision. There are common problems in most of the large enterprises:
Ideally both approach for information management ultimately aims at the same goal: to realize an enterprise wide vision to company’s business data, however, there are some differences the way both are looked at and the way one could be leveraged by the other in order to successfully realize the power of enterprise data vision. There are common problems in most of the large enterprises:
- Absence of enterprise wide data strategy
- Application based architecture
- Departmental data in siloes
- Lack of strong data governance
Presence of any of these could lead to redundant and unreliable data, which would definitely result in enterprise’s overall growth. Therefore the goal for every large enterprise remains in providing a “single version of truth” around data. Now an EDW enables this vision by centralizing data in one place which could be accessed by multiple departments or LOBs, enabling consistent and accurate version of data. Whereas MDM could be looked at as an integral part of the EDW that provides robust set of technologies and processes to materialize the single version of truth.
Let’s take an example of a bank, which has multiple products to offer to its customers. A customer can hold a a savings account, a checking account and a demat account. Without an EDW it would be impossible to bring the same customer under one identification number, thereby resulting in redundant copy of the same customer. Further, without proper MDM in place an enterprise would have to cope with fragmented view of customers. An EDW thereby is the first step when an enterprise has started consolidating its disparate data scattered in silos and MDM is the next step to ease out the effort of integration by providing common and uniform set of master data elements like customers, products, organization hierarchies etc.
The below chart describes evolution of EDW along with MDM, what falls common in between them:
This clearly underlines the important role of MDM in building successful EDW in order to mastering the data assets of an enterprise, thereby improving business outcomes.
