Depending on what you are trying to achieve, there are two main overarching approaches to supplier master data management. Here is a short reference which can quickly help you understand the differences and pros/cons of some of the different approaches.
Analytical Master Data Management (MDM) – focused on providing an MDM that can “measure the business”, downstream of the transactions and in the data warehouse supporting Business Intelligence
Operational Master Data Management (MDM) – focused on providing an MDM that can support the processes “running the business”, upstream of the business transaction, at source, where data is originally authored.
Depending on what results and benefits you are trying to achieve will determine the right approach for your organization.
If you are interested in consolidating data primarily for reporting you could be 100% satisfied with an analytical MDM solution which if we boil it down to very simplistic terms means maintaining a mapping of individual records to a common group.
An analytical MDM solution will provide you the ability to reconcile different records into one and also help to identify duplicates. At the most basic level you could send your data out to Experian or Dun & Bradstreet to do this grouping exercise for you but this is not a sustainable or cost effective long term approach.
What an analytical MDM solution will not provide you with is operational efficiencies and better data quality but it is a very lightweight approach and very simple to implement. The typical challenge here is around the governance to maintain ongoing mappings – this is where using an MDM tool versus a BI tool will help you to institutionalize that process.
Some example analytical MDM hub styles for analytical MDM are:
- Registry Style
Registry style is mainly used for identifying the duplicates [in real time]. Once the duplicates are identified, it links them together in a common group. It is a really light weight MDM Style of implementation and consolidation is not part of it. Some consider this to be operational MDM but in our view this does not operationalize master data.
- Consolidation Style
Typically in consolidation style, the master data is consolidated in Hub. It can then be synchronized back only to data warehouse. It consolidates [merges] to create single version of truth.
If you are interested in centralizing business processes around data maintenance, or to enable your shared services organization then operational MDM is the way to go. In the context of supplier master data operational MDM solutions or typically referred to as Supplier Information Management.
In order to embed the supplier data processes into the organization it’s no longer sufficient to just collect master data attributes.
The two most common Master Data Management hub styles for operational MDM are:
- Coexistence/Hybrid Style
In coexistence style, data is mastered in source systems and then synchronized with Hub. This is great example of coexisting the data in source system as well as in the Master Data Management system. It does consolidate to create single version of truth.
- Centralized Style
Mostly this shall become the choice of organizations where the Hub becomes the source of golden version of truth and mastering happens in Hub only. You turn off creating data in any of the source systems, as it tends to create duplicates. Downstream systems can always get the master data from Hub in either real time or in batch mode.
Understandably, there are different reasons why different approaches to how the hub or central system is used. In our view the most efficient long-term approach is to have a centralized style approach. It has the least amount of interface complexity and will deliver the lowest TCO and most efficient processes.
That being said, it should be noted that implementation requires the most buy in from business stakeholders as it is the most intrusive in terms of real business transformation and has the highest change management complexity of all approaches.
It’s very common during project rollouts to have a Hybrid approach as you transition to a centralized hub, especially for very large-scale deployments.