How does master data management change BI?

Master Data Management (MDM) fixes one of the biggest impediments to the use of Business Intelligence in companies. Because of the way Enteprise software packages evolved over time and got implemented in companies - like Financial Accounting first, Sales Accounting second, Warehousing and Shipping third as examples, different IDs were created and used for the same entity - product, customer, etc. Consequently, there is no Single View of a Product or a Customer and companies were unable to really make good use of the data they had in plenty. For example, my bank had my customer ID in my Bank account different from the one they created to sell me Insurance and may be Mutual Funds. So when they wanted a complete picture of what me, as a customer, bought from them over time they were out of luck because there was no automated way of putting it all together! With MDM they can now. MDM is the glue that really makes islands of BI all worth something in the end!

Peter,

Given the author's argument that "the lack of a compelling case to fully harmonize business processes" let's them continue with all the redundant systems and calling it "best practices" is hard to take from the beginning. On the contrary, the harmonization of processes is just this compelling case but there mostly seems to be a lack of willingness for fear of change, for diverging interests of various stakeholders and for steering clear of the write-off of previous investments.

Cheers,

Michael

MDM does not change BI in general terms. It is just a way to unify the business objects among different systems, thus allowing a unified vision of business data.
This is good, to exploit synergies, and bad, when the unified vision is considered as the new Gospel against specific users needs.

Very broad subject for just a blog comment, but I will raise one issue. A long time ago (1989), a very wise Professor of Anthropology at the Harvard Business School, Shoshana Zuboff, wrote a book called, "In the Age of the Smart Machine: The Future of Work and Power." At one point, she created a word, "informatating," to describe how organizations collect data as part of a process, but manage to make it infinitely more valuable by using it for other purposes. Scanner data in a grocery store or dispensed prescriptions from a pharmacy, which are gathered by third parties and re-used.

What does this have to do with BI and MDM? In our research, we have found that most knowledge workers shun BI for two reasons (not performance or ease-of-use): relevance and understanding. MDM adds nothing to address these concerns because its representational framework, a relational schema, is inadequate. It's a representation of a model. An ontology is a model. Until MDM rejects the relational model as its underlying schema, it will be unable to add the rich meaning, relationships and even reasoning that an ontology can do.

So, the point is, if you're going to go to the massive effort and expense of an MDM solution, tahke some advice from 21 years ago and "informate" the process. Make the data useful for people, not just governors and black belts.

Neil Raden
Hired Brains

MDM forces BI system designers to face the question they've always faced: Who is the customer? What is a customer? Questions that are more easily asked than answered.

Wow - I was going to say that BI systems must benefit from the "single version of the truth" that MDM purports to provide, until I read Neil's comment (above). He's surely the BI + Process expert.

So, instead, I'll pose another contextually-relevant question: if MDM doesn't help improve BI systems, then what can be done to drive superior relevance and understanding, thus enriching BI for a broader audience?

My past answer to this question would have drawn on the now popular web 2.0 techniques, including embedded BI (so it's part of an application and process), in-memory analysis (for fast, simple understanding), and mashboards (for pulling together disparate data to make information more relevant).

Neil?

Brian Gentile
Chief Executive Officer
Jaspersoft

MDM changes nothing; you are still left with massive transactional data to deal with from a BI standpoint! It’s more of an implementation technique to consolidate information and present them in a consistent and reliable manner. Why not use an SOA framework instead and expose the information as services?

The whole concept of BI is still trapped in the realm of a relational database system which itself is restrictive and that’s one reason BI hasn’t found wide acceptance among the business users.

People still end up using fancy reports driven from databases, drill downs using dimensions and measures because that’s what is available today – but no one asks the key question – for what? The basic thing that any business would benefit from is a set of KPIs which are specific to their nature of business and that’s what I feel, BI is all about. The other things are peripheral and mere implementation techniques.

Neil, I appreciated your comment and enjoyed seeing Shoshana Zuboff’s name and work invoked again. Concerning your comment:

"Until MDM rejects the relational model as its underlying schema, it will be unable to add the rich meaning, relationships and even reasoning that an ontology can do."

...I wrote a piece awhile back that touches on that notion: that The Semantic Web is something like “repository-less MDM,� in which data are linked together in meaningful ways for immediate consumption and action, rather than being delivered in a manner that requires further effort and analysis by the user.

(In light of your comment, perhaps “model-less MDM� might have been more precise.)

The Semantic Web: Repository-less MDM?
http://enterworks.wordpress.com/2010/09/03/the-semantic-web-repository-less-mdm/

Whenever I see "MDM", I cannot stop thinking that it's nothing more than covering up messes.
While leaving underlying messes, does "MDM" really make things any better? It might make things a bit easier, though.