Choosing master data management (MDM) software differs from selecting other technologies for managing enterprise data. The database market, for example, is so well established that buyers can take a relaxed view that there is no wrong choice among the top three players. Pricing and cultural fit may vary, but it is difficult to dispute that the technology is tried and tested.
That is not the case for MDM, warned Andy Hayler, CEO of London-based analyst firm The Information Difference. “There is not that level of maturity,” he said. “There are still significant differences between the technologies depending on what you try to do with them.”
For more on what governs MDM technology choices
Find out how BBC Worldwide is using product data management to boost revenue
"Big data" has brought new choices to bear, contends Andy Hayler
Data governance framework as bedrock of MDM strategy at Air France-KLM
Technology choice in MDM should be governed by a well-structured process, Hayler said. Buyers need to look at the technical constraints of their own IT environment and identify their MDM requirements in some depth. “Maybe for your company certain features are really important,” he said. “Maybe you have got very high volumes of data, or maybe hierarchy management is more complex for you.”
However, too often, he lamented, MDM software buyers are swayed either by the preferences of their dominant IT services providers or the presence of incumbent data management or enterprise software vendors.
Most MDM vendors have grown either from a background in managing product data or customer data. Although many now claim to manage master data across the enterprise, their domain history can be important because it often will suggest where their strengths lie, Hayler said.
He added that buyers could also consider a mix of MDM products, with a domain specialist for the most important function and a generalist to bring the enterprise estate together.
Mike Ferguson, managing director of analyst firm Intelligent Business Strategies in Cheshire, said multi-domain MDM products are attractive, but organisations should be wary of using them for managing complex product data domains. “Product master data is a rule unto its own,” he said. “That’s why you still see specialist vendors in that space and large vendors who have multi-domain MDM offerings but also have a separate offering for product master data. You can be caught up in the hype [of multi-domain MDM], but you need to make sure that it fits your needs.”
Outside of product and customer data, there are specialist tools for managing master data in niche markets, Ferguson noted. For example, in the investment banking industry, tools are available for handling master data relating to securities and counterparties.
If companies do need to manage master data in specialist domains, it is important to prioritise them at the start of the software evaluation process, Ferguson said. “The real challenge, before you go looking in the [technology] marketplace, is to understand which [domains] you need to manage and for what business reasons.”
Ferguson also advised buyers to be aware of the tools they will need surrounding their MDM systems and what they may already have available -- for example, tools embedded in data warehouse systems that could be used in building MDM capabilities. “It’s increasingly the case that as the MDM market has consolidated, the tools to handle the integration and consolidation of data into MDM and synchronisation out of it may be offered separately from the MDM product itself,” he said. “And the way in which those things are bundled can vary dramatically.”
The trickiest part of an MDM project can be agreeing data definitions and hierarchies across business units; such data governance programmes can cause friction between different departments. In an April interview with SearchDataManagement.co.uk, Aaron Zornes, chief research officer at The MDM Institute consultancy in the US, said that MDM software vendors have been sluggish in responding to governance requirements.
While some vendors have introduced workflow functionality that could be useful in tackling data governance, Hayler said solving the governance problem still requires a great deal of attention. “It is beginning to dawn on vendors,” he said. “A number of products have got some notion of support for data governance. It is not really a technical issue, but nonetheless they can help it along a bit.”