Navigating the Intersection of Analytics and Data Management – Part Deux

By on

by Kimberly Nevala

As analytics becomes pervasive, standard operating paradigms for data management are being challenged. In some cases new models compliment established practices while others are on a collision course.

Rounding out the list we began last week, here are five final points to ponder when thinking about how to adopt, compliment, extend or bypass your existing data governance and data management practices for an analytics-infused world.

  1. Data as a Business Byproduct or a Building Block?

There is a tendency to think of business processes/systems as something that happens to the data. The truth is a little different: data also happens to business process. In other words, the integrity and quality of data input has a germane impact on the effectiveness and efficiency of the business process itself. Data can even influence or shape the form our processes take. Workarounds for bad data very often create tortuous practices or seemingly illogical processing bypasses.

A retailer to remain anonymous once shared that they modified the process of computing the internal capacity of their trucks to account for what they knew was bad dimensional data coming in. The question was posed: what happens when good data happens to bad process? A simple example, perhaps, but one that illustrates a point played out at various levels and with broad implications across companies.

As analytics becomes pervasive, the awareness and management of data not just as an artifact but an integral component in business/system processes is critical. Careful consideration must be paid not only to what data is fed into the process but also what information and insight might be deliberately consumed or generated within the process itself.

  1. Inform or Act?

The goal of BI has always been to deliver the right information to the right person at the right time. Today, however, analytics capabilities have progressed beyond being informative to deciding what must be done next and taking action – sometimes autonomously. As a result, analytic systems must themselves be considered as an actor in key business system processes.

Leveraging analytics in this way requires mindfulness in developing both data and analytic literacy of constituents (internal and external) across the company. As well as more clarity than ever about how different cohorts behave and use (or could use) more targeted insight and information in their day-to-day work processes.

  1. Insight Found Here or “Insight Inside”?

Building on the prior point, the new analytic experience can best be described by a single word: pervasive. As analytics pervades all aspects of an organizations, insight generation and delivery no long occur in a data warehouse or delivered via reports or dashboards distinct from business processes and systems. Rather, insight is delivered at the place and time it’s needed. Up to and including embedding analytics systems into the fabric of core business processes.

Not only does this affect how we architect our data environments it requires a robust portfolio of analytic capabilities. In many cases, contrary to what big data as a term implies, this means delivering smaller doses of highly targeted information more frequently. Or analyzing all (modeling at scale) but acting small (executing on discrete events / micro-moments as events occur).

  1. Insight or Knowledge?

Variety. In the context of big data, this is used to describe the vast array of increasingly unstructured and unexploited information assets available for exploitation in today’s embarrassingly rich data ecosystem. The ability to exploit these assets for competitive, differentiating insight is pivotal to success in a big data, analytically driven world.

Analytics aside, there is an incredible amount of content (e.g. process and product documentation, FAQ, SOP) that need to be considered when enabling an information-savvy enterprise. And while the structure, format and discrete capabilities to create and curate such content is different than “data” (big or small), the overarching need for governance and management is the same.

The Chief Analytic Officer of a major health providers describes her job as “creating knowledge”. To that end, knowledge management, not just data, should be addressed in a holistic information management strategy.

  1. Information Owner or Broker?

In order to manage the incessant onslaught of data, companies are increasingly:

– Counting both internal and external (partners, customers) as parties to the data ecosystem

– Ingesting information generated by third parties over which we have limited control.

– Utilizing technologies such as data virtualization to access information on-site without timely and expensive data movement and physical integration.

– Storing and/or accessing information via the cloud vis-à-vis “cloud data syndicates” or “data marketplaces/exchanges” (private and public).

Operating in this environment requires data management teams to operate as data brokers. Rather than data owners with full control over the goods. Those who have the capability to connect multiple providers and consumers (internal and external) to maximum advantage will take the day. Does your data Uber?

In Conclusion

The data and analytic ecosystems of the future will be tightly intertwined, if not indistinguishable. What is perceived as wildly implausible today is fast becoming reality. Yes, this will require re-architecting both technical platforms and business practices. But even more so, it starts with changing the mindset regarding what is, and is not, possible or even desired. As Einstein counseled, “we can’t solve problems with the same thinking we used to create them.”

Leave a Reply