At the Insights Association NEXT conference yesterday, Andrea Jones-Rooy of New York University shared some perspectives for the questions that leaders should ask when working with data scientists. As Andrea showed in her Venn diagram, where traditional researchers have statistics and substantive expertise, data scientists add hacking skills; the two groups are more similar than dissimilar. As a result, I think these questions are useful for insights professionals and researchers in general, whether managing teams or responding to clients and corporate leaders.
“Leaders don’t realize what they’re asking for,” Andrea said. Data scientists need to think about whether something is easy or difficult, useful or not, typically starting with “easy and useful.” She shared a story of when she analyzed trends of gender representation across a workforce in different countries. The follow-up request was to do that research at the manager level to recognize managers that have more diverse teams. She decided “it was easy but not useful” since she knew as a statistician the results would correlate to that of the national workforces. “Unfortunately, I was interpreted as difficult or hard to work with. I hadn’t made it obvious to outsiders why this wasn’t a good idea.” She suggests leaders ask:
- “Is what I’m asking for difficult – and in what way?”
- “Is it likely to yield useful insights?”
- “Is there a simpler way to get similar insights?”
- “Would you do this step if you were on your own?”
- “What do you think is the most productive next step, and why?”
“Leaders don’t like the results, so the results are ignored.” Andrea commented, “I’ve seen so many companies waste so much money. I did a month-long project – the leader looked at the results, which were different from the working hypothesis, and ignored the findings. What a waste!” When leaders are faced with surprising results that run counter to their expectations, they should ask:
- “Can you walk me through how you came to this conclusion?”
- “How confident are you in these findings?”
- “What would help you be more confident?”
- “Would you be willing to implement a policy based on these results?”
“Leaders are afraid to admit they don’t understand.” In fact, leaders need to “lead by example and show vulnerability.” Unasked questions can fester and lead to long-term problems. Questions leaders should be asking:
- “Can you help me make sure I’ve got this right?”
- “Is there another way you can walk me through this?”
- “I’m not sure I see the difference between X and Y, can you clarify?”
- “Can I say this back to you to see if I’m understanding?”
Researchers (and data scientists) should make sure they are answering these unasked questions when presenting to leaders. And leaders should ask them!
Author Notes:
Jeffrey Henning
