Following our previous posts from the 8th annual MDM/DG Summit in New York, we focus today’s post on the United Technologies case study and their comments on distributed workflow and collaborative MDM. During their presentation at the MDM/DG in New York John Nicoll-Senft and Larry Keyser provided background about United Technologies Corporation (or UTC) that explained the rationale for their MDM program.
UTC is a multi-billion dollar multinational corporation with operations, to quote Nicoll-Senft, “in every country in which it is legal to do business.” Historically, their business units have been fiercely independent. This independence has translated into financial success, in 2012 UTC posted revenues exceeding $50B. However, independence has also led to, what Nicoll-Senft described as, “huge amounts of inconsistency across the business.” Inconsistency, especially in master data, hampers initiatives at the corporation to standardize processes, work instructions and analytics in supply chain/supplier management, corporate finance and tax. Addressing this data inconsistency is why MDM is on the critical path for several programs at UTC.
A good example of how complex this task can be is in Tax Technology. In order to manage consolidated legal entity reporting, the MDM/RDM team in tax technology needed to determine how UTCs 7,000 managerial entities mapped to the 1,400 legal entities and the 1,350 international tax and 550 domestic tax entities. The metadata (or reference data) for these entities is housed in multiple enterprise applications.
For Phase I, the master data came from several Hyperion Financial Management instances (managerial), GEMS (legal entities), and Corptax (international and domestic tax entities). For Phase II, master data will be added from Longview (tax planning) and Big (Real Estate). Complicating matters, these are not all one-to-one mappings. The managerial entities for a business unit will have different relationships to corporate tax, tax planning, legal entities and real estate. Pratt and Whitney, while managed as a single business unit, files taxes in every jurisdiction where operations are located.
The multidomain nature of the data requires two types of processes to address the data governance issues completely. There are the intra-domain processes, or the workflows within each domain (managerial accounting, legal entity, and tax entities) that coordinate activities between those that maintain the accuracy of the private reference data. And then there are inter-domain processes that coordinate the activities of the data stewards and owners from different domains. Consider the inter-domain relationships, or mappings (between managerial:tax, tax:legal, legal:managerial), because each ‘leg’ or ‘edge’ touches a different domain, multiple governance teams must coordinate their activities. Within the UTC corporate tax technology example, we can see examples of cross-functional and cross-organizational workflows as there are several instances where workflows span business units (BUs) or functional areas (HQ tax, BU tax, treasury).
Given the cross-functional and cross-business unit processes, for Tax Technology distributed workflow is vital. Bear in mind distribution of responsibility does not mean giving up centralized approval and accountability. All it means is that the MDM/RDM program can distribute data quality tools and processes to those who have the most “local” or tacit knowledge about the domain. Distributing workflow tasks to your business teams transforms data quality/governance into a shared responsibility. Providing entry points to the business teams gives them the means to participate in the governance processes, building engagement and political support for the MDM/RDM program.
More importantly, delegating responsibility makes it easier for the program to scale. Larry Keyser shared a humorous anecdote about the beginnings of the supply chain modernization program. “The first time we used this tool, we loaded the first 250,000 suppliers and we found 200,000 to 250,000 opportunities [for improving data quality]. It was a tsunami, a data tsunami, and the guys we made responsible and accountable for the updates were not happy.” By taking a more collaborative approach, they have moved from a tsunami of work to 30/60/90 day deliverables that make the daunting task of supply chain transformation more manageable.
For more of the UTC presentation, please view it in our Library.
By Conrad Chuang