Dan Thorpe’s session at the M2010 data mining conference was not what I expected to find at a gathering of data analysts, statisticians and data mining professionals. No mentions of second order polynomial response functions, no mathematical formulas, just a straightforward discussion with a seasoned leader about how data mining professionals – or, anyone in the professional world, really – can up their game in today’s workforce.

As Senior Director of Member Insights for SAMS Club, a membership-based grocery and service retailer, Dan’s job is to fully understand the behaviors and needs of its members to build sustained, profitable loyalty. And while there are plenty of tangible processes for achieving this, Dan’s passion is lies in the leadership skills of the people who are at the core of those processes. According to Dan, great data miners have:

1. A multidisciplined background: Dan used to look for stat degrees and a good GPA, now he looks for balance: science disciplines, for example, show experience building lab teams.
2. Both domain expertise and functional expertise: being the best modeler is great, but it’s not enough to help you grow. You also need domain expertise.
3. The ability to function as a broader team: Be able to own the bigger problem. The biggest barrier to success for data miners: “It’s not my fault.” Jump in the game, own the problems at a team level.
4. A good attitude: Just being right isn’t the answer. Be the kind of person that people want to work with. One measure of success is whether we get invited to the table at the beginning of a strategic decision vs. close to the end of it.
5. A well-crafted resume: Don’t discount this traditional job-seeking tool. It has to be written well; it has to tell a story and show growth.
6. Strategic scope: Are you helping others to succeed? Did you exhibit broader value vs. simply coming to the table as a “quant.” Think to yourself, Why am I a data miner? Do I just like numbers, or do I see beauty in the numbers, the potential to make a difference?
7. An understanding of success, not just of the analytics: Maintain perspective enough to identify the right indicators of success and keep the end goal in mind.

Dan also shared a few not-so-best practices – the hardest one of which is to avoid cutting and pasting analysis output into reports or recommendations. People have no clue what the output means, so put it in the business language and call out the Ah-ha's clearly. Because, as he quotes author and Babson College professor Tom Davenport, “How you communicate your results to decision-makers is just as important, if not more so, than getting the results themselves.”