A Day in the Life of a Chief Data Officer
Defining the value of your data
Recently I had the chance to spend some time with a group of Chief Data Officers at a conference. As a data scientist, it was great to get some insight into how the CDO role thinks about data assets and initiatives. One thing that struck me was that most of the conversation was less about technology than it was about communication. We talked very little, if at all, about the cloud, data architecture patterns, programming languages, platforms, and frameworks. We didn’t even talk all that much about data lakes or complex data integration projects. Instead the conversations were more about communications and partnership rather than technology.
The group of CDOs was very communicative and open about their challenges and desire to do a better job in several areas:
- Alignment with business initiatives and stragetic partners
- Consistent achievement of small definable goals
- Defining the strategic and financial value of data assets, so that projects can be better supported and prioritized
The goals of the business need to drive the data strategy. With good data governance and stewardship, in return the data can drive the business. Too many businesses do not recognize the value of their own data. In order to unlock this value, the data and business teams need to align and partner to set common goals and priorities. Too often, the data strategy and the business strategy can diverge. How often are teams working on multi-year data projects that do not actually provide timely and material support to the primary needs of business analysts and leaders in the organization? How often are projects worked without a clear understanding of the actual strategic goals and expected benefits of the work?
Data Life Cycle
Generation of data -> Retention of data -> Consumption of data -> New Data Assets -> Business Goals
Most business processes will generate data, usually transactional data. This is the first step in the data life cycle. Next, the data is ideally retained. The retention time may need to be adjusted to meet the needs of analysis, which is generally a much longer time than what is required for operational purposes. For many data science projects, indefinite retention of data is ideal so that patterns over time can be studied. Next, the data needs to be consumed by some analytical process which would generally restructure the data into a report, model, or other new data asset. This data asset should be aligned with the goals and objectives of end-users and their business objectives. The asset itself may restart the cycle, where the usage of the asset generates new operational data which should be retained and analyzed.
Defining these needs and aligning data projects takes communication and trust between business units. When analysts are given data access and tools to transform data as needed to make tactical and strategic decisions, new initiatives can be undertaken. This can lead to defining the value of data assets based upon the business financial goals they support. New goals can also be defined including data monetization and data-driven service personalization.
A successful analysis will generate more questions than answers
Here are some important themes to consider when you are trying to run a data-driven business.
- Good data governance, stewardship, and data literacy in the organization is key to success. Without a good awareness of the opportunity for using data to enable business processes, it can be hard to get executive level sponsorship for initiatives.
- Deriving business value from data may require not only good reporting and information availability, but also advanced ML initiatives. Staffing for both reporting and ML can be challenging as they require different skill-sets.
The amount of interest in machine learning and artificial intelligence amongst the CDOs that I spoke with was impressive. CDOs understand that you don’t have to be running a tech company to derive benefit from better offer targeting, more efficient operations, more accurate revenue forecasts, a better understanding of customer personas, and all the other potential benefits of AI initiatives.
Before you can build a good model, you need good data.
CDOs have an understanding that their organizations need to design to support the needs of the business with techology solutions and innovations. If they can fill the ranks of their staff with leaders and developers who understand that as well, success will follow.
