Tony has hit the nail right on the head, starting with the title of his article: “Demystifying MDM”.

With much fanfare and hoopla MDM chants echo everywhere, truly transcending the status of JABA (Just Another Bad Acronym (thanks F.H.)). IBM, Oracle, SAP, SAS and others swear by it, so it must be good. Many other related vendors, product and services continue “lining up”. Okay.

The article’s opening example, about customer knowledge, is on the mark – after all, in today’s personalized and consumer-oriented culture, the closer you get to your customers, the closer you are to true customer-centricity and attaining competitive Nirvana. And Tony provides sound advice: “start small” and “set up… a Center of Excellence… with both IT and business users providing guidance and support.” His valuable to-do list – about data quality, organization hierarchies, ownership/accountability policies, change management and data definition – could be expanded to include the concepts of business and IT processes definition, metadata modeling, data value ranking and ROI analysis, security and regulatory compliance…

… but wait, aren’t we simply talking about good ‘ole Data Strategy and Governance? Do we need the MDM marketing hype to pay attention to what we should have been doing all along? I guess we do.

But the story doesn’t end there. Are we talking about ALL data? In their recent whitepaper, “The Expanding Digital Universe”, March, 2007, IDC estimates that: “in organizations, unstructured data accounts for more than 80 per cent of all information; less than 10 per cent of organizational information is currently classified or ranked according to value.”

That means a plethora of information, of knowledge, explicit and tacit… in emails, text messaging, call-center conversations with customers, in forums, blogs, wikis, the media, in competitors’ literature and analysts’ reports, strategy meetings, conference calls, at the water cooler… how do we master-data-manage all that?

For the vast amounts of information being produced, an all-encompassing management approach to all data, structured and unstructured, remains a challenge. Many are tackling the issue and progress is being made. However, it is vital to pull it all together, in context, to derive an in-depth understanding of the competitive landscape and of customers’ preferences, affinities, values, desires and sentiment. Only then do we truly “demystify” MDM and get closer to the utopian “single version of the truth.”

Adriano Gonzalez
Sherman Oaks, CA
net-strategies@pacbell.net

I agree with what Tony is saying but need to point out that this problem needs to be pitched at the CEO of an organisation and perhaps even at the Board as a whole. For Master Data defines the organisation, its market, competitors, suppliers and position and role in the cycle of supply and demand.

The concept of "Process Organisation" has been bandied about a lot and some organisations have flirted with it. The fact is that Master Data Management will only be able to be maintained when an organisation has fully embraced the process organisation.

The make up of the Executive Committee of a process organisation should be end to end process owners. All supporting processes must be represented but as suppliers to these process owners.

The corporate budget, which is the tune the organisation dances to must be made available only to the process owners who in turn choose which supporting processes they wish to employ in the delivery of the business and its competitive advantages. Dead wood support will drop off, those wishing to remain engaged will have to both motivate and prove their necessity to the bottom line.

The Master Data and its definitions are not owned by the CIO, although the policies and standards will remain the CIO's responsibility. Neither will the single or multiple ERP and other system owners own the Master Data or its definitions. Such ownership belongs to these Executive process owners. The CIO is accountable to ensure that the data definitions across all integrated systems are aligned with these business definitions.

A fatal mistake would be for the Business definitions to be prescribed by business to the CIO or vice versa, but a clear understanding of what the organisations market defines (process owners), together with the constraints of the system architecture (system technical experts) and the extent of process and data abuse that has taken place over the years (Business, Business Process and System Analyst (preferably resources that have experience in uniting all of these three well developed skills)) must be jointly determined.

This is a moving target and requires specialist consultants to design Master Data and any required systematic and Business unit integration, workflow and or restructuring.

The point being that Master Data implementation serves no purpose without providing the ongoing business requirement of maintaining or in some cases improving the flexibility the Executive Committees requires for agility within the competitive environment.

Many Boards and CEOs do not have the necessary guts and strong leadership to implement the above and as a result Master Data will continue to be the "Straight Jacket" which continuously cramps the organisations delivery and ultimately its profits and goodwill.
The right consultants are also required. Rick Evans +27 73 577 1141

Here are some demystifying observations about MDM:

1. Take a look at your infrastructure, and ask yourself: Do we have One-Thing-in-One-Box (OTOB)? Are customers neatly tucked together in a single administration? Are contracts filed logically in a single storage application? If your infrastructure adheres to the OTOB principle, then you do not need MDM.

2. Suppose your information is scattered all over the information landscape, in a disorganized fashion. You can do one of two things: either you reorganize the whole thing or you search for ways to live with it. MDM fits in the second strategy.

3. If your vendor argues that you want MDM, ask yourself the "killer question": do I want more technology (for example MDM) to make my messy information system tolerable, or do I need to clean up the mess?

Well, that's all folks. I hope this demystifies MDM a little bit. I am keeping it short, because I have got work to do: cleaning up yet another large company's messy information.... I wish you the best of luck.

Stef Joosten
(full professor of information technology and business processes at the Open University of the Netherlands)

We recently went through implementing master data management at our company (Fortune 500) and our experiences are in line with this article. The focus needs to be on process.

But process only insures future Master Data is clean, and doesn't address the legacy data quality issues.

There are a wide variety of MDM solution sellers in the market offering to fix your problems.

We found many of these vendors were looking for six figure investments to help clean up our current data. They also were very poor at standardizing item information.

Their implementation times were also out of hand, weeks or months.

We found a great packaged solution from a company called Data Ladder at http://www.dataladder.com
Using their products we were able to standardize our item data and harmonize customer and vendors in a few days. I don't usually highlight a vendor publicly, but in this case we saved so much time and money I think other CIOs can learn from the experience. Legacy data cleanup can be done QUICKLY AND CHEAPLY. Just avoid being 'sold' and find a solution that makes sense.

Tony's article provides a good presentation on some of the issues relating to deploying an MDM solution as well as the more comprehensive MDM program. I address a lot of the same topics in much greater detail in my recent book, "Master Data Management." In fact, I have been very careful in the book to focus on the processes and techniques leading to a unified master view,deferring the discussion of an actual architecture approach until most of the business requirement groundwork has been laid. For more information on the book, I have put together a companion web site (www.mdmbook.com) that will contain thought leadership articles about MDM from a number of leaders in the field.

Interesting post. As usual, technology can be used to overcome data and political issues. However, your point about simply implementing an MDM solution in isolation is a good one; throw it into the fray without the requisite process improvements and data cleanup will not yield the benefits sought by the organization.