Harnessing the potential of data: A balancing act

Diligencia’s Patrick Lord recently made the opening remarks for a panel discussion on the title topic as part of the International Compliance Association’s Big Compliance Festival. In his introduction as reported below, Patrick highlights the importance of every organisation having a clear data strategy in place.

“The world is awash with data. If you think about all the data points that are available across the public domain, commercial data sources, social media and usage data from mobile devices, all of which are expanding day by day, there has never been a better time to work with data and take advantage of the opportunities and insights that come with it. Yet it is striking how many organisations are failing to derive value from data. Recent research by consulting firm McKinsey found that despite private sector organisations spending a combined total of more than US$2tn on data in a single year¹, less than 20% had achieved data success².

Why is this? Are organisations being overwhelmed by the sheer volume of data that is available? Or do they lack the capability and technology to integrate and make sense of the data they have? Perhaps it is the case that different departments or divisions with an organisation have access to different datasets with little visibility of each other’s respective activities. Having a clearly defined data strategy is the first step an organisation can take to address some of these real-life issues, and it is heartening to see that 56% of this audience³ work for organisations with such a strategy in place.

There are three main areas to consider when formulating a robust data strategy:

Risk-based approach – as a former risk consultant it is perhaps not surprising that my first port of call is always to gain a clear understanding of the risks an organisation faces. And I mean risk in its pure academic sense of not only vulnerabilities but also opportunities. Financial crime is a threat that most banks and financial institutions take seriously because they want to avoid the losses, regulatory action and reputational damage that may result; but by the same token successful use of data and technology to combat financial crime can also become a competitive advantage, a way of enhancing reputation and strengthening relationships with customers. Only by understanding the problems to be solved and the opportunities to be exploited can the foundations of data success be laid.

Resources – determining the resources required (external data, technology infrastructure, human capital) is where expectation and reality can become unstuck. A resource plan should build on and be proportionate to the risks identified, but a successful balance between volume of data, effective technology and the need for human intervention can be elusive. It is easy after all to acquire and accumulate a large data lake, but if you then need large teams of people to clean and analyse the data in order to derive any meaningful insights, then I would suggest something has gone wrong. The concept of data quality is a useful lens through which to view this problem; by strictly defining the structure, format and source of data that you require upfront and adjusting your processes and resource plan accordingly is likely to lead to a more successful outcome.

Data sharing – when talking about data it is difficult to avoid touching on issues of data protection and privacy, although it is not the intention of this panel to focus on this as we could easily spend a session on it alone. My point about data sharing is more related to organisational structure in that data and data strategy should be discussed and owned at the top level of an organisation – probably as part of a Chief Information Officer’s portfolio, although I have heard of a few examples where a Chief Data Officer role has been created. Whether you are in compliance, marketing or sales, ideally a common set of internal data resources (e.g. a single view of a customer) can be made available together with the different dimensions and analytics that make it relevant to each respective function.

I hope that helps to set the scene for this panel session, and I will now turn to my fellow speakers for their perspectives.”