Checklist for Chief Data Officer
The function of the Chief Data Officer, or CDO for short, is multifaceted. Although it is comparatively new, the job profile is changing strategically. One challenge is certainly to make the constantly growing flood of data more usable with the help of scientific methods, emphasizes Jeffrey McMillan, CDO at Morgan Stanley.
Limiting the role of the Chief Data Officer to profit maximization falls short. The CDO's position is essentially designed to achieve three key business objectives:
- drive comprehensive company growth
- Improve the efficiency of measures
- Manage risks
The idea is that all business processes and activities benefit from these three objectives. Jeffrey McMillan , CDO at Morgan Stanley, made some recommendations on this subject at the recent CDO Summit in New York. He focused on five aspects of analysis or strategic processes with regard to the handling of data.
- The data science strategy should be consistent with the corporate strategy
Good data scientists are not easy to find. But rather than spending too much time trying to find the ideal data scientist, McMillan believes something else is at least as important. Namely, that this person is savvy about business and sales issues. If no one in the company takes care to implement the CDO's recommendations, it undermines his or her position.
- Users should be encouraged to work with data visualizations
Data sets should be made available to as many employees as possible rather than left in the hands of a few. For McMillan, this approach secures vital corporate interests because it brings the data to the decision makers. "Corporate decision makers don't need algorithms. What they need is information they can use in practice."
- The framework for action
McMillan has developed a process that makes it much easier to make decisions. He refers to this framework as the "next-best action framework." This refers to a system that learns, evolves and adapts in real time. McMillan describes this process as follows: "Every single detail, about what an employee is doing or can do in his or her area of activity, is recorded into this system. This data is then compared with the employee's own expectations, historical behavior, customer behavior, market conditions and about 400 other factors. We then successively optimize the system in terms of the specific needs of the customer and the employee in question. This process brings to light a whole range of ideas that are evaluated and ranked according to a point system. For example, whether it is better to call a customer about a bad check or rather to invite them to a golf outing. And then we observe what the customer does."
- Help digital intelligence make a breakthrough
When we talk about artificial intelligence, the business value is in what is in "intelligence." That's why McMillan prefers the term "digital intelligence." "We are digitizing human understanding in a way that adds value. Ultimately, the winners will not be those who merely provide a technology. It will be the companies and institutions that have knowledge. Those who have knowledge and information will prevail in this field. Someone has to "tell" a machine or an algorithm exactly where to start. A machine doesn't learn on its own."
However complex the subject matter, McMillan exhorts his audience to make everything as simple as possible. "Rest assured, in the end, no one cares how difficult it may have been to make it that simple: Only whether it's simple."
- Opt for a holistic approach
McMillan issues a stark warning: If you don't take a holistic approach to data management, the effort will either fail or fall far short of the results a company expects. The focus should be on the most important aspects of corporate strategy.
