It became apparent in last year’s Effective IT Survey that businesses have a strong desire to sort out the quality, reliability and auditability of their data. At the start of 2010, it found, more respondents planned to deploy master data management (MDM) during the next 12 months than any other IT strategy. Gartner’s market predictions told a similar story; the analyst company predicted that the MDM market would grow 14% in 2010 to reach $1.5 billion.
However, this year’s Effective IT Survey found that actual adoption had increased only slightly, from 26.7% in 2010 to 28.8% in 2011. Clearly the demand is still there – MDM was once again the strategy that most respondents plan to adopt in the coming year – but realising that demand seems to be problematic.
Demand for MDM is being driven by its strong appeal to the CIO, argued Gartner research vice president John Radcliffe. “With MDM, CIOs can create a unified view of existing data, leading to greater enterprise agility, simplified integration and, ultimately, mproved profitability,” he said. But at the same time as predicting MDM growth, Gartner identified the issue that may explain why, among Effective IT Survey respondents at least, that growth has barely materialised. From now until 2015, it argued, two-thirds of organisations that initiate an MDM programme will struggle to demonstrate the business value.
That prediction was supported by this year’s survey: only 28.9% of MDM adopters said it had so far delivered the desired return on investment. At the same time, only 17% said that it had explicitly failed to do so. For the majority (54.1%), it was simply ‘too soon to tell’. The reason why the value of MDM projects is hard to realise, Gartner argues, is that MDM projects are typically driven by the IT department, rather than by the business itself. This is problematic because organisation-wide data management initiatives such as MDM require buy-in from all affected parties.
If MDM – or any other large data management project – is to be successful, business divisions must understand and accept which data they are responsible for, and they must understand the right way to maintain it.
This calls for a data governance framework that defines the rules and responsibilities associated with data, and for a set of metrics that spell out what the business is trying to achieve with its data management efforts. “Unless organisations take a holistic, business-driven approach to MDM, addressing governance and metrics requirements in particular, they risk having their MDM programmes fail,” said Radcliffe. “Internal politics won’t be brought under control without a governance framework, and without a metrics structure there will be no way of objectively defining what success looks like and measuring whether or not it has been achieved.”
In other words, the governance framework and the metrics must be defined by the business if they are to be useful, and must be led by the business if they are to be adopted. Another reason why large data management initiatives require leadership from the top is that employees often see the data they work with as a tool to control their working environment. “People hold information very close to their chest,” Gartner analyst Anne Lapkin told Information Age in August 2010.
All of this means that when it comes to implementing an MDM project or similar, selling the benefits to the CEO is possibly one of the CIO’s most important tasks. To do this, a causal link must be drawn between data quality and the bottom line. This is by no means a straightforward task.
MDM technology vendors have a number of familiar, hypothetical examples that they use to express the benefits of MDM tools in business terms. In December 2010, for example, data integration vendor Informatica presented these arguments in a white paper entitled What your CEO should know about master data management. Bad customer data, the document argues, can lead to missed cross-selling and up-selling opportunities; it can result in inappropriate offers being sent to customers; it can impair a business’s ability to interact with suppliers or channel partners; and it can result in poor account management.
These hypothetical examples are all undeniably possible, and are indeed articulated in terms that a CEO might appreciate. But does that mean an MDM project or data governance initiative would necessarily resolve them?
Process data management
According to Forrester Research analyst Rob Karel, data management initiatives that do not integrate with business processes are unlikely to translate to the kind of outcome that CEOs really want. “High-quality and trustworthy data sitting in some repository somewhere does not in fact increase revenue, reduce risk, improve operational efficiencies or strategically differentiate any organisation from its competitors,” he wrote in July 2010. “It’s only when this trusted data can be delivered and consumed within the most critical business processes and decisions that run your business that these business outcomes can become reality.”
To express this argument, Karel has coined the term ‘process data management’. The idea proposes that effective data management initiatives must focus on three kinds of process: “the operational processes that capture and update raw data; the processes that consume and rely on trusted data to run your business; and the stewardship processes needed to administer and maintain the data itself”. “Too often, MDM efforts only focus on data stewardship and ignore the upstream and downstream implications that impact the ability to deliver [the] anticipated MDM business benefits,” he wrote.
Karel is arguing, therefore, that data management and governance still have some way to evolve before they can be relied upon to deliver the kind of benefits that win the backing of CEOs. And yet he also holds the view that the business must own MDM and data governance initiatives if they are to succeed. (Indeed, he is preparing a report entitled It’s Time For The Business To Own Master Data Management Strategies.)
This paradox might be frustrating to hear for IT organisations hoping to resolve their data management issues through MDM. So too might Gartner’s prediction that two- thirds of MDM projects will struggle to prove their business value. And it raises a question: if proving MDM’s value is too difficult to justify adoption, will it ever have the chance to mature?