From Ready Room to C-Suite
With analytics at the heart of every data-driven business, a new data hero has emerged to deliver the enterprise from the woes of data deluge. The Chief Data Officer is a new role being added to the C-Suite, positioned as a leader, advocate and chief technologist of all things data. But what kind of challenges will a CDO actually face? Perhaps the better question is, what are the enterprise drivers that make this new role necessary?
There is many an ongoing debate in social-tech about the responsibilities and requirements of the new role. So here is our take on what a Chief Data Officers primary focus should be.
First, it's about understanding how a business uses data for productivity and competitive advantage. A good CDO needs to know how to measure and report and how to make sure that information is stored and used in the most effective way.
To use a Star Trek analogy, if the CEO is Kirk, the CDO is Spock. Someone who lives by information and understands the true nature of analysis, control groups and double-blind experiments. Someone who has the authority and ability to challenge business assumptions through data-driven storytelling.
The Prime Directive
In the enterprise, a data officer's prime directive should prohibit the CDO from directly interfering with internal development of data management systems. Their goal is to observe and drive top-down change of data architecture and technologies as they pertain to production and consumption of critical decision-making information. Leadership and alignment with business strategy are not the purview of a CDO. Many in the organization will claim that role, resulting in analysis-paralysis. The data officer should recommend, inform and challenge assumptions based on the available information; and be able to identify when there is not enough information to make an informed decision.
The goal of a CDO is to improve the productivity of an organization's knowledge workers by ensuring that they have good actionable information. This often means that a data officer will have to be the fast-fail catalyst and the Experimenter-In-Chief capable of separating innovative technologies from industry hype. Most importantly, the CDO must be able to answer the critical questions: What are we trying to measure? Why? What impact on the business will the resulting data have?
The reason for this is that Data Science is often not. Most executives invest in data analysis for the wrong reason.. they are searching for data in support of their hypothesis instead of using data to form a hypothesis. In doing so, many organizations use data to perpetuate a false consensus bias that leads to bad decisions. Instead of telling a story the data is used as supporting evidence in a trial of personal expectations.
Infinite Diversity + Combinations of Data
The data officer must understand data diversity and its causes within an organization. Enterprise data is fragmented across siloes for a reason. Different groups are trying to use the same information to answer different questions. As a result, the same data are often re-assembled in various combinations, leading to redundant data copies.
A CDO can add tremendous value by understanding how these questions may relate across various business units and whether emerging technologies can improve data synergy and reduce data copy. As often is the case, various teams that use the same data do not communicate or strive to identify common goals.
To be successful a data officer would need good listening and negotiation skills, but they would also need to be grounded in certain realities. Assume you can't get rid of dirty data or selfish users. A CDO will remind you that data drift occurs because missing values are less relevant to certain business units. So a single version of the truth may not be possible or practical. Self-service data preparation really isn't. The CDO will remind you that self-service and visualization tools require some schema or meta-data. So data curation needs to happen before prep and visualization tools can be used. And that is often the job of another team.
To Boldly Go
Most firms embroiled in analytics have complex systems, not complex problems. Complex systems are non-linear. Correlation between cause and effect is often not apparent or non-existent. Big effort does not always translate to valuable results. Outcomes cannot be predicted. The best you can do is observe and react. But the faster you react the more you learn. That means that data agility, how fast you can extract value from information and turn it into action, is pretty important.
One major challenge all data officers will face is changing the way DBA and operational staff think about data. Agility depends on an organization's ability to handle data stored in databases or file systems as well as in-motion application data. With modern architectures data is definitely going places it has not been before, like mobile devices, streaming systems and in-memory compute grids. The CDO will need to understand how information moves thru an organization and the best ways to achieve data agility without sacrificing process or introducing security risks.
In the final analysis, complex data systems cannot be managed by direct application of technology. Rather, they are improved thru observation, experimentation and in many cases, reactive data processing; that is the right application of technologies that improve data agility. That in turn leads to knowledge worker productivity and better decision making, allowing companies to expand into new markets and take advantage opportunities faster, through smarter use of data. That is the mission of the Chief Data Officer.