A while back, I ran a four-part series on this site about the notion of a formal data strategy. (Read the first installment here.) TLDR: Adopting one certainly can't hurt, but I'm not a fan because the intelligent use of data should be part and parcel to every organization's business strategy.
Today I'll discuss what place – if any – master data management (MDM) plays in such a strategy.
A tale of two organizations
Allow me to channel my inner Dickens. Consider two large, mature, international organizations with complicated systems, interfaces and data warehouses. This is where the similarities end. Organization A manages its data exceptionally well. Data quality is extremely high and it has imbued a strong culture of data governance. Throughout the organization, employees understand the importance of data.
Next up is its counterpart: Organization B. Its internal data management practices are disastrous. Few employees comprehend the data-related consequences of their actions. The words data governance have never been uttered in a meeting or casual conversation. Duplicate, inconsistent and incomplete data pervade its systems. It's a mess.
Notice how I omitted any mention of MDM in each company's description. This wasn't an accident. One organization is vastly more prepared to take advantage of its data than another. Organization A doesn't need to deploy an MDM application. For its part, Organization B needs a great deal more than just the latest MDM doohickey with its attendant bells and whistles. No application by itself – no matter how powerful – changes an organization's culture, especially one as deleterious as this example.
In the second part of my series, I described how organizations are increasingly using data for both offensive and defensive purposes. You'll get no argument from me on the benefits of big data, but make no mistake: the small stuff still matters. I have yet to hear of any organization, department, division or group that struggles with managing its internal, structured information while gleaning amazing insights from external, unstructured stuff.
Truth be told, for certain types of organizations, Hadoop (or an equivalent tool) is a far more essential component of an effective data strategy than any MDM application. Organization A falls into this bucket. For others, however, Hadoop is secondary – or even tertiary – to other, more fundamental tools such as MDM.
Depending on where your organization lies, perhaps it's beneficial to table Hadoop experiments. Generally speaking, perhaps it would benefit more from the basic blocking and tackling that relatively unsexy MDM applications can provide.
Yes, it's important to to find needles in massive haystacks – or signals in the noise – that is big data. However, all of that big data doesn't mean very much if your organization cannot simply, quickly and routinely produce a single version of the truth with regard to customers, employees, products and the like.
What say you?