At one point or another, all data in an IT manager's world in our community was related to MPE and the HP 3000. That day might be today, or it could have been last year, or in the previous century. The prospects for the future of data management are shaped by the existing design of data flow as well as business practices. Those practices define a Master Data Management plan on your migration platform as a business issue, according to MB Foster.
The company's CEO Birket Foster led a webinar on Masters of Data Management last week. "The first thing to do is look at your application portfolio," he said. That begins with a list of applications and their attributes, then fan-tails out to the sources of the data for those apps. Methods to add, change or delete, as well as where data is stored, are other elements to track.
"You want to find the code that relates to each of the screens or batch processes that deal with database items," Foster said. "You want to look at how you enforce those edits of the data."
You also want to understand the architecture of the data, he added, even when you can't control that architecture.
Foster mentioned that the databases his company is studying for its customers "are just getting retrofitted for geo-location, because it turns out if you put pin-pricks on a map, aligned to data, you can get a whole diferent picture of what's going on." Where a model of car was purchased, or mapping addresses of your patients who have had hip replacements, for example. Marketing benefits from such new data alignments.
Such retrofitting must wade into existing data structures. A data manager might benefit from having professional services to understand and control how these retrofits can be integrated.
The biggest thing in the process is "to figure out where data is held, edited, and where it's searched from," Foster said. He added that his company does an X-ray to show where data is used in apps, or what reports get fed by that data. The main message is to start with a team of people across the effort who can work together "to drive the quality of the data, to make sure it's represented across your enterprise in the same way," said Foster's Account Manager Chris Whitehead.
In addition to being parts of Master Data Management, data cleansing, completion and updating are subjects of a Gartner Group study of Customer Data Integration. List consolidation, purging of duplicates, merging records: this is the work of CDI. Adding new information to existing data -- like the legacy and founding databases at 3000 sites who're on the way to a new platform -- makes it more useful to a company.
Gartner's study denotes six levels of awareness of Master Data Management
- 1. Initial awareness of the problem, but no specifics
- 2. Developing -- firefighting mode, isolated
- 3. Defined -- Silo-level initiatives
- 4. Managed -- higher-level organizational sponsor, unified version of data
- 5. Optimizing -- a defined process, managing your data as an asset
- 6. A continuous learning process
"The thing that's changed about this over the years is that it's not only about the data that might show up in something like a retailer's cash register tape," Foster said. "Now it can be data that also shows up on your website, or a description in a brochure. Or there might be instructions to a customer on how to assemble a purchased product, like something from IKEA."
It's all master data that an IT director needs to be able to be aware of, in order to use it well. Once you've figured out where all the data is, and the workflows in your organization which use that data, then you can to creating One Version of the Truth for the data. This is Master Data Management. "You make sure that if data gets changed in one spot, it gets changed in all the others," Foster said. "How are the fields defined, and who sets those rules? Are they enforced by the programs?"
This really about helping your business, he added. "You have to put stewards in place, with good governance and good policies behind the effort in order to make yourself successful." Creating a plan for Master Data Management means consistent practices, uses a boardroom walk through for buy-in, and can drill in to look at data. The plan must be kept current over time, remembering all stakeholders. "Transparency in the plan can help with a data-driven company," Foster said. "There might be five different places where customer data is kept, and you need to track all of them. Without MDM, it makes it trickier. You have to put in some kind of mechanism in your organization to get there."