I was talking with a sales person today who wanted help to initiate a dialog with a client that we knew had an MDM “issue”.  We explored several ways we could help the client with respect to MDM, and as a result of the conversation, the sales person came to a mighty conclusion: why is there not a simple, standardized, blue print for the client to explain to how to get started, and what the next 2, 3, and 4 steps should be?

This is a good question – but I had to explain to the rep what, to me, was obvious.  Each enterprise might end up with a vision and strategy for MDM that seems to look and smell like the next guy, but that is not really important.  What is important, and now clear, is that each and every enterprise starts their journey toward MDM, from a completely different place!  As such, there is no single blue print, but several, even many.  I explained that there were large patterns emerging, but these are today not detailed enough to provide individual blue print for each firm.

There are some typical places from where large clusters of firms have begun their MDM journey; here are some of them:

• Departmental Madness: Specific business leaders, perhaps VP sales and marketing, report that business performance is poor; initiatives to improve customers service, up-sell and cross-sell, are not delivering on their promise despite significant business and IT investment in new application and business intelligence software.  Root cause analysis shows that IT costs are higher than expected in support of increased data cleansing and integration routines; and the business says it takes too long to get the data to make the right decision.  This “departmental” cluster is complicated in that any number of departments might make the leap between the symptom (poor business performance) and cause (master data quality).  So this cluster is actually many different sub-clusters,
• The Unintelligent Business, or, ‘BI gone wild’.  This cluster has the common characteristic that each business meeting, that includes leaders from different parts of the business, spend more time arguing over the data, the source of the data, or the accuracy of the reports they each use, that business decisions are often left unmade, made with poor data, or elevated to senior management who have even less idea what to do about the decision.  Generally each stake-holder has developed, over some period of time, their own data source and this source is independent, and likely not integrated to any formalized information architecture in the business.  This emerged very often when senior executives encourage internal competition for resources and management attention, between departments, or business units.
• The Sunny ERP Uplands: A mid-size firm, in the specialty chemicals industry, is trying to wrestle with a migration from many legacy and custom made business applications to a single instance ERP strategy.  This is a “me too” strategy since the firms’ competitors seem to have done this already, and the firm believe that this is a necessary step in order to remain “in the game’.  There has been little thought to how ERP supports the firms secret sauce, business competitiveness, process automation, or process innovation.  IT has no idea where the master data is; who really uses it, or why.  There seems to be lots of data everywhere, locked up in applications that do not play well with others, will be ‘sunsetted’ over the next couple of years anyway, and bottom line – business and IT do not have a good working relationship.  IT is distrustful of the business – they do not think that the business really knows what it needs.
• The Kaleidoscope: A large, global financial services firm is trying to migrate hundreds of legacy and ERP applications across several regions. around the glob, over the next 5 or more years.  There are regulatory issues to cope with as there is an additional global BI strategy that supports financial reporting.  The final selection of two large ERP alternative vendor is about to be made; the business seems to be ignorant of the fact that compromises in each case will lead to different parts fo the business being serviced with innovative or best in class capabilities. 
• The Engineer: A large, European automotive parts supplier is being hit with two major issues: the need to reduce substantially its time to market for new parts; and increasingly demand and more complex requests from customers and partners for product related data.  The firm has invested in Product Lifecycle Management and this was thought to be the silver bullet to their product development needs; this did not prove to be the case.  While engineering received some good functionality to meet their design needs, the firm seems to be have been over-sold by the PLM vendor that implied that PLM supports the end-to-end lifecycle of the product.  The firm is now wrestling with how to manage information across the business, across PLM, ERP, SCM, CRM, Procurement, and other systems – including trading partners.

You might see aspects of each of these clusters in your business – and that just proves the point.  Every firm will understand and agree that there is value in having consistent master data (we can even agree that need the principles of MDM in the home – every tried trying to change channel without a good, clean program guide?), but the place from which every firm starts the journey is different, and infinitely complex. 

I have only looked at the vendor’s SCM, analytics, and MDM technologies.  In the SCM market the vendor sports some pretty good, even best of breed, capability, that heralds back to the companies founding.  Agilisys was formed out of the spin-off of the SCP engine that was part of SCT Adage, a process ERP/manufacturing vendor.  Agilisys, originally a niche SCP vendor, became Infor.  Later acquisitions have led to the consumption of a discrete best of breed SCP engine also, originally known as Mercia Software (a vendor I had interviewed with way back in time, though I turned them down in order to move to the US).  Other acquisitions have led to some other interesting offerings, though some not neatly aligned to the two SCP offerings already mentioned.  Thru-Put Technologies was an old “drum, buffer, rope” engine (think The Goal) designed to optimize discrete product scheduling, that had been acquired by MAPICS, that was itself acquired by SSA.  So Infor now owns that IP but alas it remains focused in-plant, and not yet applied to distribution. 

Analytics, or more precisely, Performance Management (that is, the embedding of analytics and business intelligence capability inside the business process) and MDM remain challenges.  Admittedly these would be challenges for any vendor that has acquired so many different applications each with their own architecture.  Infor has an interesting approach to supporting analytics (see the note above) but certainly has changed, for the better, its views on MDM.  When I started talking to Infor about MDM (June 2005) MDM was a lower priority, even following the foundation of a SOA strategy.  Back then I said that this approach looked odd; how could an architect settle on an SOA strategy that promised cross department/application integration and composition, without some notion of an end-to-end (interoperable data model?  Not 12 months later, MDM had moved up the priority list at Infor, and even now, is a much higher priority for the vendor.  This is no criticism of the vendor; this question comes up many, many times in end user conversations.  Infor was, and is, reflective of how our mutual customers view this technology.  Thankfully Infor is now more able to participate in the conversation with its users about how it can help in this area.

Bottom Line: MDM, as part of a methodical and broad EIM strategy, should precede and/or be part of, any broad SOA strategy.  See When SOA Breaks, What Then? and SCM Vendors Support Master Data Management as a Precursor to SOA.