97 Things About Ethics Everyone in Data Science Should Know

Chapter 98. The Gray Line

Phil Broadbent

Senior Manager, GCX Analytics, eBay

I stood in the amphitheater-shaped room presenting results from the latest machine learning model we had implemented. It would optimize prices at an online retailer that was embracing the novelty of data science in order to drive revenue. I looked up to face my accuser.

We’d had modest success to that point and were proposing a new metric to measure the impact moving forward. We avoided most of the usual political turmoil that exists in bigger companies by maintaining a relatively flat organization, both in hierarchy and in operation. Folks with good ideas that delivered results were consistently given room to make decisions and own strategic direction.

One drawback to this structure is the ability of folks with different agendas to derail progress that others make. Such was the case on this day, as a leader of a marketing team (which had been tasked with driving up demand and was struggling to do so) expressed doubt in the metric we were proposing. They accused my team and me of “misleading” the organization by showing the results in the way we were. Most of the folks in the room were not quantitatively minded and didn’t see the attack for what it was: a political maneuver designed to position one group higher at the expense of another.

As a reader, you may react by questioning the details (they are ...