Practical Pragmatics and Data Monetization

I attended two Chief Data Officer meetings recently, and two topics permeated the discussions.  First, the challenge of how do firms properly manage the increasing amounts of data that’s created?  This includes data governance, quality, technology, operational concerns, and more broadly, the necessary work around good data custodial needs.

Second, how do firms monetize their data?  This spawned a much wider debate on what monetization even means.  Can firms wrangle the massive amounts of data they have collected to make more informed decisions and cut costs?  Or does monetization instead mean being able to bring differentiation to clients and add value by creating new revenue opportunities?  Further, what data matters most when it comes to either one of these interpretations?

It is obvious that the starting point is conducting data governance in order to even have the capabilities to monetize data.  In short, you need to align your data to your business priorities.

When it comes to business priorities, serving the customer is an obvious one.  But what about enabling internal processes across different business units?  And there are the needs of the enterprise in generating profits while ensuring compliance with local and global rules and regulations.

One could argue that, in serving the needs of the customer, this should naturally ensure the enablement of internal processes and result in the expected benefits to the enterprise.  At the same time, one could take the view that serving the needs of the enterprise is the leading vision, as this will necessarily drive a focus on the customer and the processes needed to achieve that vision.

But these are vision statements, as opposed to tools we can use to manage data.

To monetize data, it’s to understand where you and your data sit within the larger social system of your organization and the wider system of financial services. This is where pragmatics come into play.  Pragmatics is a linguistic discipline aimed at understanding intended meaning, and how context contributes to meaning, especially in social interaction.

In data, we tend to worry about semantics and the meaning of data, and assume that signs in the data can be used to determine the context.  But without pragmatics, the formal semantics that guide computer science efforts in ML, AI and NLP tend to not consider original intent.  The purpose or reason any data may be stored, from the perspective of the community that created the data, has influence over the actual meaning of that data.  For example, in social interactions, a listener works to derive the intended meaning of a speaker by considering their history, tone, body language, etc.

How does this impact data monetization and data management?  As data users, there is a tendency to assume that all data is known and meaning is certain.  But meaning, intent and context are not stored with data. Therefore, the data user can easily infer different, sometimes erroneous, meanings when they come from a different community of practice and have a bias toward another context.

So, in the pursuit of data monetization, the pragmatic concern of where the data comes from and the intent not stored with the data need to be considered in tandem with where your data is going, along with the intent and context of the customers the data is expected to serve.  If they are not aligned, you will likely find that you are not creating the value you intend for the end customer or your enterprise.