Data leadership is still a new career area, with a history that stretches back only to the early 2000s. Even Chief Data Officers are still trying to find the best swim lane. In the most mature organizations, they report directly to the CEO; in others to the CIO or CTO. They may even sit under Operations, Marketing, or Finance. 

Another clear sign of data leadership’s short history is that many of us started our careers doing something else. I’m an accountant by trade. I fell into data through being the guy who was always looking for clever ways to display information in Excel, then working to understand the next step beyond that. Some data leaders have been database administrators, managing mainframe systems and then being promoted up the organization through the technology structure. IT people come to data with their own set of skills, including architecture, modeling, or performance management. Data leaders also include ex-medical doctors, educators, and aerospace engineers.

At the same time, most data leadership degree programs are still in nascent stages, with a three-day bootcamp here and an online certificate program there. So what will the next wave of data leaders need to present themselves as the complete package? As we’re seeing a convergence of the scope data leadership needs to oversee, I would argue for several skills:

  1. Data architecture and modeling. As a data leader you need a firm understanding of architecture and data modeling concepts. Which platforms and tools should you use to analyze and store data, and how will you work with your data architect to arrive at the ideal platform? You also need to be conversant in building a data model and choosing things like attributes, rules, and entity types that your model should include. 
     
  2. Data warehousing. Because you’re likely to be working in Azure, AWS, or Google Cloud managing terabytes of data and millions (or even billions) of rows of data on a regular basis, you should have some working knowledge of data warehousing and how this maps to cloud-based services. Managing the expense of cloud services is another challenge, so transparency to who is running or refreshing jobs at what frequency is key. 
     
  3. Cloud-based analytics. If your company has been around for more than ten years it’s likely that there are digital disruptors in your competitive set. They are likely to be using analytical modeling tools and data models that live in the cloud, and that are shared and stored there as well. To be a data leader, you need to think lean and agile, and to move toward using cloud-based analytics to become more data-informed in your behaviors. 
     
  4. Experience with much larger (and external) data sets. Data leadership means not only understanding the architecture and source of data, but how it gets transformed, loaded, shaped and adjusted. These days that includes bringing in and augmenting data from third-party business sources, from government sources, even from weather sources. And you have to understand how that can all fit together, whether you work in real estate or aviation or population health.
     
  5. Data Visualization. Once the information has been brought together, data leaders need to understand the visualization side as well. How do you and your teams build dashboards that show meaningful relationships and help you to tell a compelling story with data? Being able to demonstrate effective output is crucial to the credibility of a data leader. 
     
  6. Data ROI. Data leaders now have to have an understanding of ROI as well. Success is therefore about much more than saying to a business colleague, “Here’s some data that we’ve prepared and here’s a dashboard we’ve built, off you go.” It’s about considering the return on the piece of analytics work that you’ve done and how do you measure that? 
     
  7. Data Fluency. A lot of organizations talk about data literacy. We prefer data fluency because to be data literate means that you would have been illiterate before, and we don’t like that negative connotation. Fluency is about data leaders’ ability to understand information as a second language, as Gartner said a few years ago. 
     
  8. Data Ethics. To be a data leader you must understand data ethics as well as the different geographic connotations. The rules on data privacy in the UK are different to what we have in America, APAC, or even Germany. As the volume of available data grows, data leaders need to strike a balance between privatization and personalization. Even though people want everything customized to their personal experiences, they also want the privacy to not have to share all their data. And just because technology will allow you to do amazing things these days, you have to judge what you will do against what you legally can do and what is the right thing to do. The world moves faster than the laws that are in place, but then they tend to balance up, so it always pays to be ahead of the curve on ethical issues.
     
  9. Business Acumen. The final part of data leadership these days is understanding the business – the stakeholders, what the business does, how it makes money, how it’s compliant, or how it reduces costs or improves productivity. This wasn't really a requirement before. A business lead would simply send a data leader some data and request that they transform it. Industry was irrelevant. Now data leaders need to have a much more rounded view. 


Don’t let this laundry list of skills overwhelm you. To become an impactful data leader you don’t have to be an expert in everything, but like any T-shaped person you need to bring just enough knowledge across many areas to offer a broad-brush view of the world to your company. I understood my business first, how we made money, and how our P&L worked. Data skills came later. But we may be approaching a point where that equation is inverting.