#290: Always Be Learning
In this episode, Tim Wilson, Val Kroll, and Spotify product manager/data scientist Mårten Schultzberg discuss the limits of focusing solely on win rates in experimentation and introduce a broader "learning rate" metric that captures wins, regressions (avoiding bad outcomes), and neutral (no‑effect) results. Mårten explains how Spotify’s experimentation framework defines and measures learning, emphasizing the value of early detection of harmful changes and the insight gained from well‑powered null results. The conversation also covers cultural shifts needed to adopt this metric and how to balance experiment types to avoid gaming the system. Listeners walk away with practical guidance on redefining success in A/B testing beyond just positive lifts.
#289: The Imperative of Developing Business Acumen
In episode #289 the hosts discuss the essential role of business acumen for data and analytics professionals, defining it as both a grasp of general business fundamentals (finance, marketing, P&L) and deep knowledge of one’s own organization and industry context....