Developing a successful master data management (MDM) strategy starts with identifying broken business processes that can be fixed with improved data management, according to analysts.
MDM tends to come across like an infrastructure project or middleware -- something that IT would sponsor, according to Dan Power, president of Hub Solution Designs Inc., an MDM consulting firm. But placing sponsorship with IT misses the point, he said.
The line of business needs to sponsor the project because it can identify the business value the data holds and how a single view of that data can affect the bottom line.
"IT should be involved, but it shouldn't drive this," Power said. "It's really a business paradigm -- changing the way the business thinks about data."
Getting started with an MDM project
Master data is all the data elements that are broadly used and shared across departmental boundaries. MDM is the discipline for ensuring the consistency of an organization's data, enabling a single view of product or customer data.
"For us, it is definitely not a piece of software," said Ted Friedman, vice president, distinguished analyst with Stamford, Conn.-based Gartner Inc. "It is a holistic discipline -- process, architecture and people."
In contrast, an ERP system isn't designed to master data -- it's designed to fulfill processes, according to Rob Karel, principal analyst with Cambridge, Mass.-based Forrester Research.
Also, different SAP applications have their own set of data models, and ways of managing customers, according to Ravi Shankar, senior director of product marketing for Siperian, an MDM software vendor. In turn, data warehouses are used more for reporting purposes, Shankar said.
Finding the people whose jobs are affected by the data is the best way to identify those broken processes and thus get started with an MDM project.
"Find the business pain, and find out what their problems are, and what better, more accessible, more secure customer information and product information would do for them," Power said.
He added that organizations should start building the MDM business case with involvement from the three different levels of the organization -- the C-suite, the business owners (such as the vice president of marketing), and the line-of-business people.
Ideally, companies should try to recruit people who have a good balance of the process and data views, Friedman said. Those individuals should be high enough in the organization to have the ability to make changes, yet low enough to be intimately familiar with the problems that poor master data can cause in the organization, he added.
If the company has already invested in a data governance organization -- people from business and IT who represent the business needs and the technology that supports organizational data -- use those stewards to do the analysis, Karel said. Ask such questions as: What are the processes most in need of the data? What are the systems providing that data? Does assessment of the quality of that data improve consistency?
If the company doesn't have a formal data governance organization -- for instance, if a project manager or enterprise architect is trying to evangelize the project -- it should start by defining the priorities. If a line of business can't articulate a negative business impact around the data, Karel said, there probably isn't a reason to pursue MDM there.
"MDM is a great place where the squeaky wheel should really get the grease," he said.
In reality, what people want is a single view of the data, and that can't be completed until there is a strong foundation of clean, accurate data, according to Power. Getting different parts of the business to agree on definitions of the data can be one of the hardest parts of the project.
Because it's unlikely that a company will ever get the entire business to agree on such a broad question as what a customer is, a better strategy is to put out a definition and have people respond with those aspects that work for them and those that don't.
The scope of the project -- which master data domains a company will tackle -- is one of the most important pieces to start with, Friedman said.
Instead of looking at MDM as a holistic, cross-enterprise strategy, look at it from the bottom up and take a multi-step, multi-phased approach, Karel said. Identify specific business processes and functions for the most-improved opportunities around data management, he advised, and quantify or qualify the value of improving that particular process.
What's the ROI of an MDM project?
Fulfilling compliance initiatives, as well as reducing costs, is a common business driver for MDM projects. For instance, rectifying duplicate customer data can reduce the amount of money spent on mailings.
The company also gets a true understanding of lifetime customer value, according to Aaron Zornes, founder and chief research officer of the MDM Institute. For example, an insurance company could have three different stores of customer information, depending on what the customer purchased. MDM can increase cross-sell opportunities.
Depending on the scope of the project, an organization can derive benefits in as little as six months, Friedman said.
Avoiding pitfalls and mistakes on MDM projects
Organizations should not assume going in that the right answer is going to be consolidating all master data into a single repository at one end of the spectrum, Friedman said. There are other ways to integrate and synchronize data.
In turn, because MDM is a cross-organizational effort, there has to be some funding model associated with it that engages the different parts of the organization and makes them feel that they have some skin in the game, he said. Make sure one individual doesn't hold much more political clout or power in the initiative than any other individual.
Above all, remember, don't lead with the technology.
"There's some great technology out there that can really help support and deliver," Karel said. "But it's not going to be the main [catalyst] of MDM."