The Federal Reserve has hired its first Chief Data Officer (CDO). Micheline Casey is joining NIH’s Eric Green, FCC’s Greg Elin, and a small but growing band of federal CDOs. It seems that a position that could not have existed five years ago is becoming nearly compulsory today.
Last year, Harvard Business Review exhorted its readers that their “C-Suite Needs a Chief Data Officer.” In May of this year, Data Management Association President Peter Aiken told Information Management that Chief Data Officers were “more vital than ever.” Federal Computer Week listed the arguments for and against CDOs. IBM’s Kenneth Stockman also explored this concept earlier this year.
But how exactly does a CDO fit in—not only within the IT shops, but within the larger organization as well? And is this a position like the CIO, new but here-to-stay? Or is it more transitional, the way webmasters were in the late ‘90s and early ‘00s, but eventually saw their roles decentralized?
Data is everywhere
Data is like water, an understanding we reinforce with numerous aquatic metaphors: we are awash in data, it flows, leaks, and is subject to bottlenecks. One could say we have data, data everywhere, but that we are the fish, swimming in an ocean of data, and we don’t know that we’re wet. One of the key tasks of a CDO, then, is not only to manage and direct the flood of data, but also to help employees understand their data environment.
If only for marketing purposes, data touches every industry.
Like the private sector, the public sector is beginning to understand and benefit from the data at its disposal. City after city after city (after city after city) is putting data stores online for residents to use, and the federal government has a clearinghouse for data. As but one example, Google recently released a feature that gives the nutritional information of various food items based on a USDA dataset.
Though the value of data is clear, the means by which companies, governments, and individuals can derive that value are still being developed. And many people who populate the C-suite have not operated in data-rich environments. For them, a CDO may be a necessary addition to their ranks.
What could a CDO do?
Some federal CIOs and CDOs believe that the Chief Data Officer’s most important responsibilities are actually inward facing: to help ensure the interoperability data; to establish data sharing; to educate employees about the data they create, and how it affects and is affected by data created elsewhere within the agency; and finally, to integrate data streams into decision trees, institutionalizing data-driven decisions.
These responsibilities can be broken into two basic categories: technology and culture. Of the two, technology may be the easier one to address. Agencies need a firm understanding of the mechanisms by which their organizations create, store, distribute, analyze and present data in order for a CDO to help ensure the interoperability of that data. Already, governments at all levels are creating dashboards, pushing real-time data feeds, and realizing the benefits of open data.
But it is perhaps the cultural responsibilities of the CDO that are both more important and more difficult to discharge. In every layer of an office there are staff members not currently using data for day-to-day decision making: leaders who have learned to make decisions in the absence of data; managers who never had data generation or analysis in their portfolios; and front-line employees who feel overwhelmed already by their responsibilities and bristle at the notion that they need to layer yet another task onto their workload. It will take more than code and hardware to establish regimes of data sharing; to educate employees; and to institutionalize data-driven decisions.
Because the charge for CDOs will touch nearly every worker within an agency, from desk officers to top leadership, it is imperative that everyone understand the value of data to the mission and see the benefits. One point of agreement among experts in this field is the need for CDOs to incorporate every relevant stage of the data lifecycle, from generation of raw data to presentation of data analysis, at every level within an agency.
CDO success metric
Another point of agreement was that for many agencies, the CDO should be a temporary position, one that has a specific success metric: does everyone in the agency understand where to find and how to use the data at their disposal?
But, as one CDO explained, data is so essential to the mission of many agencies that to silo data into one office and place a CDO in charge of it would be tantamount to making the CDO an overseer for the entire agency. There would be no decision of any importance that the CDO would not influence, if not determine.
To ensure that does not happen, it is incumbent that current and future government leaders have a sophisticated understanding of how data works both in general, and how it operates within their agency in particular. Imparting that understanding should be the chief responsibility of the Chief Data Officer.
Departments and agencies that feel the need to establish CDO positions are admitting their failure to either fully utilize their Chief Enterprise Architect, position their CEA in the organization where they can do the most good, or have a CEA who can fully execute on the intended role as defined in Clinger Cohen (including life cycle management). The critical failure of most EA’s I’ve observed in the federal public sector is the lack of an actionable model that embodies both the documentary requirements identified in The Common Approach to Federal Enterprise Architecture coupled with reverse-engineered database and application code models, together with well articulated requirements models tied to underlying statutes (if applicable). Such models are indeed achievable and can be supported through less than astronomically-priced tools such as Sparx Software’s Enterprise Architect 10, used within the U.S. EPA’s Office of Water PMO. The best aspects of such an approach allow IT governance to immediately see the traceability impact of changes to the models in response to changing business needs, delivering much needed agility during a period of faster-than-normal organizational and budgetary change.
By coupling and integrating the documentary EA requirements with such models, CEA’s can provide a clear, workable bridge between their strategic EA functions and the tactical transitional implementations such architectures plan for. Moreover, they provide all the data expertise needed while giving project management offices valuable head starts in project initiation and planning activities.
[Disclaimer: this is a personal opinion that does not represent the official view of the Environmental Protection Agency.]