At Bristol-Myers Squibb, I have the privilege of working for a company singularly focused on our mission to discover, develop, and deliver innovative medicines that help patients prevail over serious diseases. Accurate, high quality, and trustworthy data is central to our work in R&D, manufacturing, sales and marketing, and corporate functions. In IT, we strive to make sure the right data is available to the right audience at the right time with the right quality and controls to advance our company’s mission. With the digital data and analytic transformation that is pervasive across the health care industry, as an IT and a data professional, there has never been a more exciting time than now to transform how we manage, protect, and consume data to help patients prevail over serious diseases.
The digital data and analytic transformation is not unique to health care. Everywhere you turn, in industry after industry, the focus is on digital and analytic transformation with companies in a race to become the digital enterprise powered by machine learning and AI. This transformation thirsts for trusted good quality data. Yet the one common theme, in my conversations with IT, analytics, and business leaders across industries, is the persistent dissatisfaction on the state of data in the modern enterprise. There is no disagreement on the aspirations of treating data as an asset and a fuel for the modern enterprise. Yet almost all enterprises suffer from the weight of legacy data infrastructure, dysfunctional data stewardship and poor rate of return on organizational investments in data management.
So what is my solution?
I believe it is data governance 2.0, a pragmatic, relentless, self-sustaining data governance aided by machine-assisted data stewardship. I define data governance 2.0 as the combination of people, process, and technology that precisely articulates the data domains and assets that are critical to the enterprise (high risk and/or high value), defines the baseline of where the enterprise is today in managing the data (data ownership, data quality, data readiness), defines the target state of where the organization needs to be, orchestrates pragmatic ownership and asset management processes that efficiently fits in the organizational structure and culture, relentlessly monitors utilization and value, and course corrects without dogma when needed. This data governance 2.0 should use algorithmic automations, machine learning, and AI to reduce the organizational burden and bureaucracy so that human involvement in data governance shifts from mundane data stewardship tasks to qualitative action directed by the “machine.”
I recognize data governance is not new and perhaps it is the most overused phrase in the annals of data management. But there is no getting around the fact that unless the modern enterprise establishes a bedrock of good data governance, the edifices of digital and analytics transformation will erode and dissipate like the statue of Ozymandias looming over the rubbles in the desert. The time has come to leverage the analytic and AI advancements of today to reboot data governance and elevate it to the same level of importance as good financial governance.
So what are the key considerations for establishing this data governance 2.0 foundation? I plan to explore this through conversations with CDOs, industry leaders, peers, and practitioners.
As the first in that series I had a chance to discuss the topic of data governance with Jason Fishbain, the chief data officer (CDO) at the University of Wisconsin at Madison. Fishbain is a passionate proponent of pragmatic data governance striving to build a strong data foundation at the University of Wisconsin at Madison. Synthesized below are the key takeaways from that discussion.
Data governance tied to strategic business objectives
Successful data governance efforts must be linked to business objectives. In his efforts at University of Wisconsin-Madison, Fishbain tied the need for good data governance closely to the educational analytic needs required to achieve the university’s strategic business objectives ranging from student recruitment to graduation goals. This enabled him to create rapid buy-in among the academic leaders on the need to own and manage the data effectively at the source. According to Fishbain, a chief data officer must be a strategic leader and a data evangelist deftly navigating the organizational structure to consistently and relentlessly advance the data governance goals. He focuses on informal and formal outreach to academic leaders to understand their priorities and see how he can enable them with the right quality data.
It is all about the outcome, be flexible on the governance model
Organizations often get bogged down in debates on data ownership and data governance models. In our discussion on how to define the right data governance models, Fishbain pointed out that in his experience, effective organizations eschew fidelity to organizational models in favor of a pragmatic selection of a model or models that best enable the outcome. Effective CDOs show the willingness to lead or to serve, be a COE or be a data stewardship unit all in the cause of enterprise data outcomes. If a certain department has the necessary skills, resources and the willingness to manage the data, then he believes the CDO should support them with standards, best practices and tools. If other areas lack this expertise, then the CDO should offer data management as a service.
Data on the state of data is key
According to Fishbain, a CDO must define and publish a few but relevant metrics on the “state of the data” to the organizational leaders tying them as much as possible to the attainment of strategic business objectives. As Fishbain puts it, if the CDO does not have data on the state of the data, then how can you shine a spotlight and mobilize organizational action? As in any good business metrics and KPI, these “state of the data” metrics must be limited, focused, and action-oriented.
Business process changes and technology shifts are opportunities for a data strategy refresh
Fishbain astutely observed that new capability deployments—be at a new CRM or ERP implementation or a redesign of a business process, are perfect opportunities for a CDO to advocate for a relook at the associated data strategy. Is the data ownership clear? What are the data quality objectives? What are the data consumption aspirations? These are all questions to ask at the launch of a new business process or technology initiative to call attention to the fact that in today’s enterprise, a lack of focus on data strategy is a surefire way to undercut the expected ROI of most capability investments.
In conclusion, there are pragmatic approaches to data governance which will allow the practitioners of its art and science—the CDOs and data leaders to steer the enterprise towards a future where effective data governance is the default than the exception.
This article is published as part of the IDG Contributor Network. Want to Join?