Personalized By Design: How to unlock long term customer value with data
summary
Focuses on designing data-driven personalization for Headspace membership and scaling to 100 million users, arguing that recommendations must deliver measurable business and user impact not just algorithmic accuracy. Cites examples: Spotify Discover Weekly, Amazon cross-sell, TikTok’s 200-day build and ~190 TB of profile data (processing equivalent of 55 billion photos daily, now >2 billion users), and Netflix where 75% of viewing comes from recommendations and thumbnails are personalized (1.8s attention, 82% focus). Recommends designer–data scientist collaboration and rules-based systems when data is scarce.
AI-written summary · about
chapters (14)
- 0:00Introduction and welcome
- 1:26Meet the speaker
- 5:07Personalization in the modern world
- 7:13The human role in personalization
- 10:01Spotify’s recommendation experience
- 12:27TikTok’s algorithmic innovation
- 16:20Business impact of recommendations
- 18:31Netflix’s data-driven growth
- 21:59Defining personalization and AI
- 25:29Building dynamic experiences
- 30:27Key questions for product teams
- 31:10Netflix’s personalized rows
- 34:23Key takeaways and summary
- 37:12Closing and Q&A
related talks
28mCrafting Hyper-Relevant User Experiences
34mA Human Touch to Consumer Experience Powered by Data
24mHow Spotify built an audience framework that lasts
24mHow to Target UX to Different Psychographics
38mDon't Talk Data to Me
31mWhy data science and UX research should be best friends
55mPlanning For Personalization: Tips for Dev, Design, & IA
22mThe 3 Rules Behind the Most Successful Products