Sitewide
RSS Feed:
|
By: Judith Hurwitz, CEO, Hurwitz & Associates Published: 18th October 2007 Copyright Hurwitz & Associates © 2007 |
Ok, so you never thought that Master Data Management (MDM) was cool. Well I just returned from IBM's Information On Demand conference and found that IBM has gotten MDM fever in a big way. In fact, they were handling out cool MDM buttons. Now, MDM has been around for a long time. Customers are always trying to get to a "single view of the customer". But because there are so many sources of data that are stove-piped and disconnected across departments, it has been nearly impossible. One of the approaches that was popular a few years ago was Enterprise Information Integration (EII). I hereby declare EII dead! What is taking its place is true MDM.
Now that MDM is starting to mature it has the potential to become the authoritative way to manage the single trusted view of the customer. I got some valuable insights into MDM from Dan Wolfson, a Distinguished Engineer and MDM expert at IBM. His view is that MDM has to be constructed as a hub. Through this federated view it is possible to understand the context of data. I really like this concept and think that if we are indeed going to move to a model where there is a single view of the customer it will be because companies can have a single way to manage their information about customers, products, services, and the like.
Let me digress for a minute and tell you a little about IBM's new approach to MDM. The company has come up with a strange name for it: multi-form MDM. Despite this name, I actually think they are on to something. The idea behind this is that rather than having a different platform for each type of MDM focused application, there is a single platform that can be used no matter what type of approach is at issue. For example, there are customer centric MDM applications that focus on the details about customer, location, relationships, etc. Another application might focus on products such as a company's product portfolio, billing details, sales territories, and the like.
There is a lot more to say about this MDM hub approach but that will have to come later. Check out our monthly newsletter. I will probably write a more in-depth article about that later.
But as I promised, here are five cool things about MDM:
1. A Master Data Management platform can help avoid a thousand versions of the "truth"
2 . If implemented from a holistic perspective, MDM can actually solve problems
3. MDM can actually become an information integration standard
4. MDM could become the lynch pin between metadata, semantic web, and registry/repository
5. MDM is cool because it really matters to the business
We are no longer accepting comments against this item. We suggest contacting the author directly.
19th October 2007: 'Andy Hayler' said:
Judith appears to be unaware that certain vendors (e.g. Kalido, Orchestria Networks) have been championing the "multi form" approach to MDM long before IBM did, and indeed when IBM were proclaiming the exact opposite was the "correct" approach, which they were doing just over a year ago. It is nice that IBM have now come around to an argument that their executives were villifyng when I presented it at the CDI/MDM institute conference less than two years ago.
Moreover she may like to read the following article, published in 2004, regarding the originality of her "EII is dead" claim:
http://www.intelligententerprise.com/channels/integration/showArticle.jhtml?articleID=23901932
25th October 2007: 'Tim Paydos' said:
Well geez, Andy, don't mince words with us. How do you really feel?
In reality, IBM has been arguing for the Multiform MDM approach since the fall of 2006, starting with the 2006 Information On Demand conference. Actually, they may well have been on the approach well before then, as evidenced by their acquisition of 2 best of breed players in the space, Websphere Product Center and Websphere Customer Center, nee, Trigo and DWL.
It appears to me that IBM's planned release of its MDM Server simply demonstrates it commitment and ability to deliver on what it has been saying all along.
25th October 2007: 'Andy Hayler' said:
Well Tim, after I spoke at the March 2006 CDI/MDM conference in San Francisco I recall an IBM exec arguing vehemently with me that the idea of a generic appraoch to MDM repositories was nonsense. Given that one implementaton I was involved with dealt with 350 different types of master data, I disagreed. When I pressed him on how many was, in his view, ideal (one for product, one for customer, one for how many others?) he claimed that "seven" was the correct number of domain-specific hubs (don't ask; he did not elaborate).
I am very happy that IBM has changed tack and realised that enterprise customers desire a consistent approach to managing master data of all types, whether it be customer, product, asset, organisation, employee or whatever. The alternative is a new form of stovepipe architecture that it seems to me that customers will resist.
Hence I am actually very much in favour of the new approach. Forgive me if I seemed to be relishing the moment rather.
19th October 2007: 'Cliff Longman, Kalido' said:
Finally! IBM has come round to the realization that the D in MDM refers to ALL types of data.
Thank goodness we were spared the horrors of GIM (geographical Information Management), FIM (financial Information Management) and HIM (HR information management) when we skipped straight to MDM.
Wait a minute! That's curious, when did the I turn into an M? Did we just downgrade Product Information Management to Master Data Management?
The acronyms seem to be changing faster than the data at the moment!
23rd October 2007: 'Dan Wolfson' said:
I think that there are really three dimensions we need to explore: 1) covering multiple domains of master data (and the relationships between them) 2) Acknowledging and supporting the different styles of use - operational transactions, collaborative authoring and analytical insight and 3) that there really are different implementation patterns - from registry to co-existence to transactional hub - that we have seen customers implement.
The important thing is to provide a way to quickly get started in a manner that addresses the most important pain points and can then be expanded over time without disrupting what has already been implemented. So we see some customers perhaps starting with an operational usage of Party data using a co-existence implementation style (still providing SOA services) and then perhaps extending that with hub transactions in a subsequent phase followed by a broadening out of the managed domains to include Account information. There are many ways to get started - but once started on the path, we need to give implementers the freedom and ability to incrementally extend the value of master data throughout their enterprise.
The messages above were all contributed by IT-Director.com readers. Whilst we take care to remove any posts deemed inappropriate, we can take no responsibility for these comments. If you would like a comment removed please contact our editorial team.
Published by: IT Analysis Communications Ltd.
T: +44 (0)190 888 0760 | F: +44 (0)190 888 0761