MDM Goes Beyond the Data Warehouse

“The Briefing Room” with Rick Sherman and Information Builders 11/15/2011 4pm
November 15, 2011
The Power of RAM: How In-Memory Changes Everything (DM Radio)
January 10, 2012
Show all

MDM Goes Beyond the Data Warehouse

articles_inside_analysisEnterprises are awash with data from customers, suppliers, employees and their operational systems. Most enterprises have data warehousing (DW) or business intelligence (BI) programs, which sometimes have been operating for many years. The DW/BI programs frequently do not provide the consistent information needed by the business because of multiple and often inconsistent lists of customers, prospects, employees, suppliers and products. Master data management (MDM) is the initiative that is needed to address the problem of inconsistent lists or dimensions.

The reality is that for many years, whether people realized it or not, the DW has served as the default MDM repository. This happened because the EDW had to reconcile and produce a master list of data for every data subject area that the business needs for performing enterprise analytics. Years before the term MDM was coined, MDM was referred to as reference data management. But DW programs have fallen short of providing effective MDM solutions for several reasons.

A Data Warehouse Does not Alter Source Data

A DW supports downstream BI reporting and analytics but not upstream operational processes and applications such as enterprise resource planning (ERP) systems. This is not an inhibitor to supporting BI and reporting, but it does become a problem when it is expected to support operational processes that require MDM, i.e., creating a master list to support ERP systems. The underlying assumption by many in IT is that the ERP systems should manage their own data, including master data, since the ERP system was where the data originated in the enterprise. Many data warehousing practitioners follow the commandment: “Thou shall not alter source data.” Without being able to change the data to create the “golden record,” effective MDM processes just cannot be achieved.

But a funny thing happened on the way to delivering the single version of the truth via ERP systems – people used more than one of them, and each one had its own version of master data. Maybe an enterprise had products from different ERP vendors or maybe they implemented the same ERP system multiple times in their enterprise specific to different business groups.

Regardless, the end result was that the ERP systems needed an MDM repository to accomplish an enterprise role. That has led ERP vendors to start building their own MDM solutions specific to their products and for other software firms to build standalone solutions.

It is More than Just Consolidating Data

Data is gathered from many data sources and integrated into a DW where it is further processed to prepare it for BI. The DW integration processes are very good at getting the data as consistent as possible within the constraints of the quality of data it receives. MDM, on the other hand, needs more robust and extensive processes that gather, cleanse, conform, reconcile, validate, authorize and publish the agreed upon master data. These processes demand more extensive tools than just extract, transform and load (ETL) tools that are typically used to load a DW.

Generally one of the first MDM initiatives undertaken by an enterprise is creating a single view of the customer. This priority is driven by the significant potential business ROI from achieving a comprehensive and consistent view of the customers. The key integration process to customer-focused MDM is resolving difference in customers’ names, either a person or a business, and their associated addresses. Name and address matching requires specialized processes and products typically not used in a DW.

Tools are not Sufficient

Another constraint to successfully implementing an MDM solution using a DW has been the assumption that technology was going to solve the MDM problems without significantly including people and processes (and of course politics) in the equation. Data governance, with the business taking ownership of the data and IT becoming its custodian, is needed to successfully design, deploy and sustain an MDM solution.

Businesspeople need to help IT identify master data, define it and assist in determining what the resultant master list should be. Managing master data needs business commitment if it is to be treated as the corporate asset that it is. It is important to note that this inhibitor is not systemic to any technological approach, but rather needs to be included in whatever solution is developed.

MDM requires that the data governance processes are part of an enterprise information management initiative to actively update and maintain the data definitions that the business creates. Data governance is an enabling process rather, but too often falls short because its deliverables are not incorporated into information management.

Getting Started

Since data warehouses and associated processes have fallen short of providing MDM, it must be time to buy a MDM product, right?

Although buying versus building may be the most cost-effective approach, let’s step back before you buy a MDM product. You should determine what is needed before you buy. Failing to understand why past efforts have come up short is a sure recipe to failure. You might deploy a new MDM product, but still not achieve the business results you want. Even worse, you might be creating yet another data silo – moving you either further from a single version of the truth.

The two ingredients to success are a comprehensive approach to data integration and enterprise data management for your enterprise.

An enterprise data integration platform addresses the first inhibitor previously discussed by enabling two-way integration between the operational and analytical solutions, using the appropriate techniques that are needed to integrate the various enterprise applications and the DW. By implementing tactical standalone data integration tools in the past, we constrained each solution built. Typically, a DW uses ETL while enterprise applications use web services, which limited the ability of the analytical and operation systems to exchange and integrate their data

Enterprise data management, particularly data governance, is essential to begin any MDM solution and stay in operation. No technology is going to eliminate the human element of these solutions.

Should you buy a MDM product? Maybe, but before you do make sure that a comprehensive enterprise data integration platform with a supporting information management program are in place to ensure success.

3 Comments

  1. A DW can handle downstream BI revealing and statistics but not upstream business procedures and programs such as business source preparing ERP systems.

  2. Confirmed methods and best methods are used to style our information factory solutions

Leave a Reply

Your email address will not be published. Required fields are marked *