Andrew White

I had a peach of a call today (actually Tuesday last week) with a customer!  This was a mid-to large firm (over $1bn) that described itself as mature (on its second large scale ERP roll out) in terms of IT capability.  I was speaking with the IT side – not uncommon of course but notable since so much of MDM is still business led.  The primary question was, “how can we build the business case for MDM so that the business sees the value?”  This is not an uncommon question either – but the client had a really good “situation” that made this all the more harder to answer.  The situation is that, as a mature IT shop, they have spent numerous years building out all manner of business applications to meet each specific business request, and along with it, the associated data (and master data).  As such, over time, more and more of the IT budget has been tied to “connecting the dots”.  IT is so good at this (so the client said) that the business users have little or no idea how bad the situation is; all the cleaning up, alignment, synchronization, and sharing takes place “behind the scenes”.  

I asked, “if you ask the business, how good and clean and timely is the data in terms of supporting decision making (business intelligence) and decision taking (business applications)?” and I was told, “It’s mostly likely that each department will respond that the data is “pretty much OK”.  As such, there is little interest in the business in worrying about MDM!  

This of course is a common situation – but this call just seemed to bring a laser like focus on this issue for me.  IT spends too much money on “integration” and has masked the real problem from the business such much, that IT cannot effectively sell the need for change.  

We then explored how in this situation the business is really governing data poorly, and duplicating effort, since each user (in each department) is governing their data for their use.  Obviously one saving for the business is that this effort would be rationalized, centralized, and thus reduced.  At the same time, in so doing this, IT would cut its spend on “integration” since over time some of the spaghetti (point to point) integration would be replaced with simpler hub and spoke integration through to an MDM repository.  

Lastly, the same client asked another common question: “OK, so we talk of MDM as if it is new, whereas we know the problem is not new – but how have users been “solving” this problem before MDM technologies came along?”  The subtext to this question was, “Do I need to consider a separate, stand alone MDM technology to enable my MDM discipline, or can my ERP system do the work?”

Well, like all good consultants, and the answer is somewhere between, “yes”, and “no”.  Surely firms have been coping with the lack of “single view” for years – this we know.  We know, for example, that some of the largest firms on the planet have used redundant copies of ERP to act as federation stores for ERP oriented master data.  But, the reality is that it is not MANDATORY that an MDM technology is needed; but it could help.  ERP (whatever that means…wink-wink) was designed to a) provide single view of (financial) transaction, and b) to meet the needs of the ERP application users.  ERP was never designed to be part of the information or integration infrastructure; it has little interest in worrying about the quality of master data outside its purview.  ERP is about homogenous environments – why would it case about what is essentially a problem in heterogeneous environments!  

Anyway, my colleges John Radcliffe and Michael Smith wrote a note, published August 2008, entitled, “Creating the Business Case for Master Data Management”.

I had a great day today.  But don’t get me to talk about my “universal communications” offering at home.  My Cable/TV/Phone service went down today and I continue to wait for the engineer!  O the wonders of “convergence” in the telco business.  Now I have one point of failure in the house.  Great.  Happy days.