Most product teams think they have an onboarding problem. In reality, they often have a systems problem. From misleading activation metrics to misaligned incentives, over-optimized funnels, and AI experiences that personalize too late (or not at all), many products are accidentally designed to create activity instead of meaningful user value. In this keynote, Sanjana Arun draws on her experience leading AI-driven ranking, experimentation, and optimization systems at eBay to unpack how product decisions shape user behavior at scale — often in ways teams don’t realize. Through real examples and live journey analysis, attendees will explore: - Why activation metrics can create false confidence - How experimentation frameworks sometimes optimize for the wrong outcomes - Where personalization and AI can reduce friction (and where they create more) - How to diagnose hidden drop-off points between signup and “aha” - Practical ways to shorten time-to-value without overwhelming users