In this podcast episode, former Netflix VP of product Gibson Biddle shares valuable insights about Netflix's A/B testing and what it reveals about user behavior. Surprisingly, he found that delivering DVDs faster didn't significantly boost retention, high movie ratings didn’t necessarily lead to higher enjoyment, and demographics were often poor predictors of viewing preferences. Instead, past viewing habits emerged as the strongest predictor of future choices. Biddle explains how Netflix prioritizes user enjoyment in its algorithms, even at the cost of potential revenue. The conversation also explores the ethical implications of autoplay and how it might clash with users' long-term preferences. This episode emphasizes the significance of observing actual user behavior rather than relying on stated preferences, while also addressing the delicate balance between user satisfaction and business objectives.
Sign in to continue reading, translating and more.
Continue