By: Ike Ononogbu, Managing Partner, InforData Consulting
Published: 9th November 2010
Copyright InforData Consulting © 2010
In the recent past the topic has been broached in all
data-related events: SAP TechEd 2010 in Berlin and Las Vegas,
Informatica World 2010 in Washington DC, Data Management 2010 in
London. And this trend is likely to continue. Organisers just
can't get enough of it, and rightly so.
Picture this typical scenario.
An insight into the business has exposed a worrying picture. Core
business entities—Master Data—reside in many applications
resulting in multiple representations of events. On the backdrop
of this, an IT-driven business strategy has been drawn up. Now,
how will this strategy be implemented?
Step in Master Data Management (MDM). The primary reason businesses look towards MDM solutions in their bid to solve this issue is because it offers one "single version of the truth" for Master Data which includes: Customer, Vendor, Product and Employee. Furthermore, centralising core business entities ensures data can be viewed and used company-wide and, more importantly, decisions are made based on the same data set. In effect, MDM allows for accurate reporting, operational efficiency and effective decision making.
To have a reliable MDM capability, three vital processes ought to be implemented: Data Profiling, Data Integration and Data Quality. These steps are aptly referred to as 'The Pillars of Master Data Management'.
This process involves statistically examining data available in existing data sources. By profiling your source data your business can have a better understanding of data patterns and formats. This understanding will pave the way for smooth data integration.
In a lot of companies, data is stored in different formats and places. Maintaining similar data in different locations gives rise to different business units interpreting data, though similar in requirement, differently. To achieve your goal you have to amalgamate data coming from these disparate sources into a single repository.
For data in MDM to serve its purpose, it has to be consistent, accurate and valid. To achieve this, data has to be cleansed and validated. In effect, Data Quality ensures data stored reflects the true nature of the business.
It is worth pointing out that all phases are equally important, though the amount of work, time and resources invested may vary from phase to phase, and no one process should be underestimated.
In the final analysis, for MDM to be successful, like any well executed business strategy, IT has to be an integral part of business. The seamless fusion of IT and Business will maximise the value, in business terms, of the company.
Posted: 9th November 2010 | By Ed Wrazen :
One of the fundamental building blocks to MDM is really understanding the business processes and the detailed activities that interact with data. What process creates the data? What processes use/update that data? What rules are inherent? Which process utilises which rules? What exception handling is performed when a rule is violated..and so on. This gives a foundation for planning and implementing the appropriate solution. Data Profiling is far more than just determining patterns and formats for data integration. This is really the basic. Data profiling should be a collaborative process conducted with business teams, information analysts, data stewards to provide the contextual understanding of the data with the business process. This will give much deeper undersatanding of business rules, data gaps, remediation processes, standardization, matching and linking of master data attributes and a raft of other considerations. Without Business and IT working together, MDM will never realize the benefits and is more likely to fail.
Posted: 11th November 2010 | By Andy Hayler :
I would add that data governance is a key "pillar" to master data management. Without the business ownership of data,and the processes in place to resolve disputes about the definitions and structure of master data, and MDM initiative will struggle. In my experience it is importance that such programs are business-led, or at the least jointly led by business and IT. IT departments alone cannot successfully push through the business changes needed for a successful MDM program.
Posted: 11th November 2010 | By Ravi Shankar :
These requirements are spot-on as you might have heard at Informatica World. The fundamental technology requirements for a successful MDM implementation are data integration and data quality. In the order you suggested, we recommend to our customers that they first perform a data value assessment by profiling the data. The power of data integration is being able to access any data, in any format, and in any latency. Last point - data quality should be integrated, not bolted on, within the MDM lifecycle. For an MDM project success (not just tech requirements), data governance is key as Andy Hayler points out.
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.
We automatically stop accepting comments 180 days after a post is published. If you would like to know more about this subject, please contact us and we'll try to help.
Published by: IT Analysis Communications Ltd.
T: +44 (0)190 888 0760 | F: +44 (0)190 888 0761