One theme I briefly touched on in my previous post was the desire for consistency and synchronization of shared master data for the community of consuming business processes and applications. However, in our experience, from a business process perspective, synchronization goes beyond timeliness or currency of the data associated with any particular data domain. Yet in retrospect, one of the most common approaches to designing and deploying master data management projects has focused on implementing a single data domain at a time, concentrating on the consolidation (or as I often term it, “dump”) of data into the master repository.
From a project perspective, the creation of a master data domain demonstrates progress, although the creation of a master data index for any domain prior to understanding the downstream usage scenarios may lead to some curious design decisions that are sincerely regretted somewhere down the road. And in fact I have seen other contractors almost come to blows over the design and architecture used for managing both master domain hierarchies and intra- and cross-domain relationships – do these live within the index, the domain, or are they to be managed as separate data resources?
The confusion over the design is a byproduct of the “single domain” approach. Businesses do not view their master domains as silos, so why attempt to design and build them that way? Alternatively, our business processes reflect a blended use of data pulled from a collection of master domains. Customers purchase products, products are built from item-list parts, the items are sourced from a variety of vendors, etc. And these business processes are not limited to a single operational environment. There may be many processes for engaging prospects, or customer satisfaction, or product performance analysis, or vendor/supplier engagement and management.
In other words, your organization’s use of master data has to be rationalized along the different dimensions of utilization, no matter what domains are referenced and no matter the original context. A more mature approach to master data design looks at consistency and accuracy in shared data creation and use across multiple sources, for multiple domains, used by all applications, and synchronized across (and even external to) the enterprise. Interestingly, this concept was nicely discussed at the recent Informatica Analyst Event, at which the MDM team presented the concept of “Universal MDM.”
The presentation highlighted the same ideas I have discussed in this blog post: Universal access to a variety of master data domains, universal deployment (with seamless integration across platforms, or even between cloud-based and on premise infrastructures), integrated within universal solutions across multiple business applications that use multiple domains, or isolated within specific applications (such as synchronizing master customer data between an in-premise master repository and a cloud-based CRM application such as salesforce.com). In essence, this approach enables a conceptual “federation” of master data environments that, if implemented wisely, can provide transparency when making master data available.
To hear more details on the topic of this blog series, you can listen to the replay of this Webinar: “Experts Share How to Launch an MDM Program Quickly & Successfully.”