Andrew White

This book (The Data Asset: How Smart Companies Govern Their Data for Business Success) arrived yesterday – and I had to read it overnight!  What does that tell you about my sleeping habits?  All kidding aside, I did read rather quickly Tony Fisher’s new book, sent to me by DataFlux (many thanks).  I found the title and overview very interesting; hence I just had to read it.  I had a few observations.

The main point is that the book had a good title, but to say that the book is all about, “data governance, data quality, and data management” confused me.  I got to page 9 and I thought to myself, “this book is about MDM more than it is anything else” and then a Eureka moment hit me – which I will explore in a moment.  Anyway, the book focuses on what data governance is, and how data quality helps, and how data management practices support the other two.  It is a good book, and one that concludes (by way of its governance maturity model) that MDM is a critical part of the overall strategy that is required – for the most mature levels of maturity (I agree).

I particularly liked the maturity model that implied varies stages of governance and it is obvious that many firms struggle to cope with governance.  Using a maturity model can help you figure out where you are, which is the first step in figuring out what to do next.  I also liked the strong and clear link between business driver and governance; this was clear from the start of the book and remained visible throughout.  Also I found the dialog on the roles within the business and IT as it relates to different levels of maturity to be very good also. 

I found the narrative relating to business rules very timely, and intriguing.  Though probably obvious to DataFlux, and maybe data quality buffs in general, the challenge with “business rules” might be widely understood, but as they relate to MDM, is only just heating up.  Many users that get serious with MDM discover that business rules, currently embedded all over the enterprise, have to be treated (in many cases) as if they were master data – in that many have to be externalized from applications in order that consistency of rule can be established.  This is not easy when you have a lot of legacy applications.  This topic will get some more attention in the market place I am sure.

It was not clear to me why “ERP” and “CRM” were highlighted as the main culprits of application silo.  This is way too over simplified and runs the risk of appearing “not applicable” to a wider audience.  The first introduction of the background of these application silos should have explored the universe of stuff that is out there, not just ERP and CRM.

I did not find the SOA and MDM connection – towards the end of the book – adequate in that it did not explain, other than high level concepts, how the two efforts work together.  I myself have struggled with this one, though I did get to compile an entire presentation on just this topic at a recent Gartner event so I know that there is a lot more to talk about here.  I guess that because DataFlux is primarily a data quality vendor I cant criticize them for not filling out this part of the narrative.

DataFlux  missed an opportunity when they talked about the process side of MDM…er….governance.  They explain how process (page 143) includes discover, design, enable, maintain, and archive.  They did not explore – they very nearly did – the idea that master data experiences a life cycle and that by understanding the lifecycle, you get to expose (discover) all that you need to know in order to design, enable, maintain etc.  The problem is that when you take a life cycle approach you understand that there are dependencies – they get exposed immediately and it turns out that these are more important to MDM (data governance) than the quality of the data at any moment in time.  We can solve “data quality today” pretty easily – you and I do it all the time for our own purposes, but as soon as we pass the data on, the real difficulties arise.  This for me was a gap – a missed chance.  

The one item that did not make sense to me was the use of “business process automation” as the ultimate stage on governance maturity.  DataFlux describes BPA (page 90): “It automates process to automate decision making, optimize processes, and reduce costs”.  For me – this is not the “end game”.  Automated decision making, when it can be achieve, sounds like a rule.  Some decisions should not be automated.  Processes can and should be automated, and/or optimized, though many processes are by design, non-repetitive (so cannot be automated) even though they can be efficiently followed.  And running a business, or process, at the lost effective cost should be standard practice.  So I am not sure what BPA really means.  Sounds like some attempt at not saying “peak performance” which would have been better.  The bottom line, MDM is the corner stone whatever the name used (I agree).

Now for the Eureka moment.  There I was, late last night, reading a book on “data governance, data quality, and data management” and I realized that I could have been in the 1990s.  I could have been in the 1980s – just.  Technically I could have been in the 1970s but there were just not that many computers or resulting silos to worry about.  There were, of course, mountains of paperwork that suffered the same theoretical issues.  And by page 9 I figured that this book was really about MDM, and how governance, data quality, and data management, help users achieve MDM.  The point was – there is no such thing as data governance!  Data governance is like saying; you need to love your wife.  Of course I know I need to do that, but what does that mean?  What are the specifics?  MDM is a specific – it implies governance, it needs some data quality, and it is all about a kind of data that is managed.  So MDM is much more meaningful, focused, and specific.  Saying “you need data governance” is just not what the patient needs to hear.  So that was it – page 9 – and I was “moved”.

Anyway, it’s a good book.  I would read it with MDM in mind whenever you see “data governance” and you should be in good stead.  Of course, before you respond, there is other data – non master data that needs to be governed.  But I would always counter: everything hangs off the DNA, the source code.  Without that [being governed], the rest wont fly.  Then I came full circle: the book was about governance and data quality in general, not “just” MDM, even though 90% of the book was focused on “governance and data quality in association with MDM”.  I gave it 8 out of 10.

Submit a customer case study to win the Gartner MDM Excellence Award 2009!  Go here: http://www.gartner.com/it/page.jsp?id=851612&tab=special&params=pg,zzz.html

1 Comment »

Category: Data Quality Governance MDM     Tags: Data Quality, Governance, MDM

Share