10 Critical Myths and Realities of Master Data Management
Prevalent myths surrounding MDM alongside an explanation of the realities.
You know how they say those who don't know history are doomed to repeat it? I think anyone who's been around in IT more than 10 years will tell you that idiom is easily verifiable in technology, where "history" arguably only spans a fat 40, 50 years or so.
Sometimes, it actually works out and people learn from the past and create smarter offerings. Witness the difference between ASP and SaaS. And then again, sometimes it doesn't. Just talk to anyone who's undergone several ERP upgrades.
Right now, master data management is at that inevitable "Y" in the road between repeating and rectifying the mistakes of the past, contends Data Doghouse blogger and Athena IT Solutions founder Rick Sherman. And the key to whether or not MDM succeeds depends on understanding why enterprise data warehouses failed.
Enterprise data warehousing was supposed to integrate the data into a single, reliable version of various types of information - customer data, product data, whatever the business needed. Sherman writes:
The reality is that for many years, whether people realized it or not, the Enterprise Data Warehouse (EDW) has served as the default MDM repository. This happened because the EDW has to reconcile and produce a master list of data for every data subject area that the business needs for performing enterprise analytics. Most of my customers referred to this as reference data management years before the term MDM was coined.
He broadly explains how EDW achieved this through integration, a unique key for each master data row and deduplication of the data. In a follow-up post, Sherman explains why EDW failed at master data and what should be done differently with MDM if organizations don't want to create yet another data silo.
The bad news is the last step's a doozy.
First, you're going to need enterprise data integration, Sherman writes. EDW failed at master data in part because it only supported data "downstream." IT made the mistake of entrusting the "upstream" data to ERP because originally, there was only supposed to be one ERP system. Even Generation Y can tell you how that played out.
Shockingly, (insert eye-roll) this two-system approach lead to two unreconciled data sets, one for analytical systems and one for operational systems. An EDI platform can help you tackle this problem, Sherman says:
An EDI platform addresses the first inhibitor discussed above by enabling two-way integration between your operational and analytical solutions using the appropriate technique, such as SOA, EAI, EII or ETL, that is needed to integrate the various enterprise applications and EDW. Typically, an EDW used ETL and the enterprise applications used EAI technology, which limited the ability of the analytical and operation systems to exchange and integrate their data.
Second, you're going to need to start managing data across the enterprise, or, to put it into a three-letter acronym, you're going to need EDM (Enterprise Data Management), he writes.
Don't be fooled by the three capital letters; this is something you do, not something you buy, and it includes hefty disciplines such as "data quality" and the ever-popular (again with the eye roll) "data governance."
Plus, as David Loshin, president of Knowledge Integrity, Inc., recently pointed out, MDM isn't a silver bullet for all your data problems. You've still got to look at the business process if you want MDM to help you achieve business goals such as cross-selling or upselling.
EDM, particularly data governance, is essential for any MDM solution to get started and stay in operation. No technology is going to eliminate the human element of these solutions.
Alas, if recent history teaches IT nothing else, it's the truth of that last statement.