
Data science and shifting roles

This CPO regrets that product management exists | Tom Verrilli (CPO of Whatnot)
Topic 9 of 20
Published Aug 2, 2026 · Clip 41:17–44:51 (3:33)
Tom has empathy for data scientists now reviewing 'half-assed' analyses from non-experts, arguing part of the confusion stems from chronic underinvestment in data engineering, structures, and labeling. The best data scientists respond by hardening tracking and attribution so misunderstandings are harder, and he notes AI doesn't absolve anyone from owning the analysis. He also argues 'fewer for the same output' doesn't mean fewer people in macro — better systems fuel faster growth — and says hybrid 'tech lead'/'EM-light' roles running small incubation teams are trending up as the cost of trying things falls.