THE DATA ADMINISTRATION NEWSLETTER – TDAN.com
ROBERT S. SEINER – PUBLISHER

Introduction

In my previous article - "Practical Data Stewardship Structures", I talked about the Data Stewardship Program and how it could be implemented from an organization structure standpoint. This article attempts to focus on the emerging role of the Chief Data Officer and its futuristic trends over the next five years and the challenges that would need to be overcome for the role to be successful and to establish a successful data stewardship and data governance program.

The Emerging Role

Data Stewardship, as defined by Robert Seiner, involves "formalization of accountability for the management of the data assets. Data stewards have been around for a while. The traditional data stewards were responsible for collecting data, and converting it into a format suitable for the servers to consume it, and keeping the data for the systems they are stewarding up to date in the database. The Data Steward was basically responsible for overseeing establishment of data management standards, procedures, and accountability for data governed within their portfolio. But in this information age, it is envisioned that this role will change drastically. A major portion of the responsibilities would still exist but the context of it will definitely broaden in the upcoming future. There would be greater focus on Data Governance which has an implicit top-down focus on establishing higher order organizational practices, policies and guidelines for data management processes. In the next 5 years, I visualize companies possessing a vision of managing data as a strategic asset and not as an IT service falling under two categories:

(1) Traditional: These kinds of companies will continue to have the conventional CIO role with the Chief Data Steward playing a role that is very steeped in IT and more reactive then proactive. It might also become common to have the Director or VP of Business Intelligence and Data Warehousing to be asked upon to act as the Chief Data Steward for the organization. The organizations in this group would most probably found to be in Stage 1 (Uncertainty) or Stage 2 (Awakening) of the Information Quality Management Maturity (IQMM) grid outlined by Larry English.

(2) Visionary: Companies falling in this category would be more of the futuristic type that would appoint a Chief Data Steward or Chief Data Officer as a peer to the traditional CIO and be part of the C-level executive staff reporting directly to the CEO (see Figure 1 below). The reasoning behind this would be that it would be overwhelming for one individual to assume all the roles. The typical CIO is focused on driving efficiency of operations, consolidation of technologies and cost-cutting. In the midst of all these, it becomes difficult for this individual to have a razor-sharp focus on data asset utilization. The future CDO will in fact help set the strategy for pursuing business opportunities. Companies like Yahoo which are quintessential information age companies would fall in this category. "It's an industry first, and I predict within a few years, most companies will have the position of chief data officer," said Usama Fayyad, who wears that hat at Yahoo Inc. Organizations belonging to this group would either be in Stage 3 (Enlightenment), Stage 4 (Wisdom), or Stage 5 (Certainty) of the IQMM grid.

Key Challenges

The Chief Data Officer will face a number of challenges as he or she tries to establish the second and third generation quality systems for the organization.

(1) Uniqueness of Skill set and Personality: The role is not for someone steeped in technical knowledge nor is it for a business person who's a technophobe, either. The individual must possesse a unique combination of business, technology and diplomatic skills. The role of the CDO must be empowered by the organization to make decisions regarding data, resolving conflicts across disparate groups and establishing enterprise standards on the use of data. The CDO's challenge would be to thoroughly understand the politics within their organizations and have the insight on how to navigate around those challenges.

(2) Ensure Visibility of Data Governance Council: The CDO leads the Data Governance Council (Tom Redman in his article on "Data Quality Governance" calls this the Data Council) is charged with defining and executing the data quality policy at the highest level. In short, the CDO's job is to provide visible leadership by clearly articulating the business purposes and strategy for the data quality effort. The biggest challenge here would be to keep senior management's commitment to quality at a level that's consistent and effective there by raising the visibility and importance of corporate data as a "product".

(3) Volume of Information: One of the foremost challenges that the CDO would face is to figure out a way to deal with the volume of data that is collected and processing it within a short time interval, reacting to it and getting it to an actionable form to the business community.

(4) Effective Supplier management: In most modern organizations, most of the data comes from suppliers and it becomes almost impossible to find and correct errors downstream. The CDO's challenge would be to make sure that a program is established for managing suppliers. This would include defining the methodology for selecting suppliers, and ensuring that these suppliers understand what is expected, measuring performance against these expectations, and making improvements to close gaps. The CDO's organization should make it that the data quality measurements that are made by suppliers are regularly communicated to the customers.

(5) Consistency of Vision: The CDO is responsible for guiding the journey of data quality and must not treat it as a destination or project. The CDO's challenge would be to figure out innovative ways to evangelize the adoption of DQM measures internally and externally with partners and will oversee the Net Present Value (NPV) Analysis for various data quality initiatives and seek appropriate funding from the business sponsors. The focus on business purpose and strategy and weaving that into the vision is important to help focus the effort, in effect enabling managers to figure out, for themselves which data and which dimensions of data quality are most important.

(6) Privacy: Managing the data assets and enforcing privacy policies would be another challenge that would need to be addressed. Investment would have to be made in research activities that discover new ways of transforming data into something that is not decodable, and yet in a form that still preserves the data's statistical value. Data as it pertains to audience behavior and habits, online buying patterns, advertising click-through, searches and medical records would fall in this category. Algorithms would need to be designed that give results from which you can never infer anything about a particular individual, but that give you a prediction of what may be happening in the future.

(7) Policies and Procedures: In order to establish a second generation data governance system, the CDO would have to take the organization out of the reactive mode and establish policies and procedures that let the data governance council function as a well-oiled and pro-active organization. The policies outlined by the Data Governance Council should address the roles played by IT and the business community and issues such as data ownership, data sharing, and data privacy. The CDO would also help in formulating the Data Governance Charter, Definition and Prioritization of activities, establishing the Data Governance Rules of Order, the Standard Documents and Forms and detailed roles and responsibilities.

Conclusion

In the next five years, it is envisioned that some companies would in fact get forced to remove some responsibilities from the chief information officer, creating a new role of chief data officer to manage the their data and ensure its security due to the separation of duties required by legislation such as the US Sarbanes-Oxley corporate governance law. The creation of a data governance team would not be considered complete without the appointment of a leader, someone like a chief data officer.