I'm just back from participating in a Chicago roundtable discussion on “Attracting and Developing Talent in Data and Analytics.” The event was run by the Executive Club of Chicago, and brought together 150 CIO/CMO level attendees across a broad spectrum of industry and academic institutions. I joined panelists Tom O'Toole from United and Eric Sigurdson from Russell Reynolds, as well as moderator Jeanne Harris from Accenture, in a lively discussion on finding and nurturing data scientists.
While we discussed a wide range of topics, these three main themes stuck with me:
Hunting unicorns. Searching for top analytic talent is often described as hunting unicorns, however our consensus was that this notion is hokum. While there is a heavy demand for data scientists (as I’ve blogged before), your approach should be methodical as with any other discipline. Think about how to build out your team with complementary skills. Still, there does appear to be a shortage in intermediate leadership positions. These are the people who understand enough about the business and the numbers to make the right connections between need and ideas, and they are able to clearly communicate the meaning of your team's analysis and why you should believe it.
Teaching old dogs. Those with a numerical bent who want to get on the analytics bandwagon should be encouraged to retrain and hone their skills. It is never too late to make the transition.
Fostering talent. We discussed how critical it is to align teams with the business, since better analysis comes from context and domain knowledge. It's vital to challenge and offer exposure to more junior staff, including giving them opportunities to present and learn. Variety is also very important. Inquisitive minds tend to be easily bored.
The panel discussion gave me food for thought as I work to build out our analytic talent here at FICO. If you have any insight from your own experience in attracting and developing top data scientists, feel free to share it here on the blog, or check out a few photos of the event (see if you can spot me).