Spotify explains why it doesn't use Bayesian A/B testing on its 800M-MAU product
generativist · x · 2026-09-13
Spotify published a thorough engineering blog post, "Why Spotify Is Not Using Bayesian A/B Testing," explaining the methodology choices behind experimentation at a household-name product with 800M MAU. Eric Seufert highlighted it as an incisive read: while most entries in this debate rehash well-worn arguments, this one stands out for its depth and for grounding the analysis in a concrete, large-scale context. The key takeaway is that methodology choice is highly context-dependent and doesn't generalize well.
More from Research
- MIT team reports in Science how 3D folding of synthetic gene circuits shapes expression — anshulkundaje · 2026-09-13
- Upstream gene dominance: supercoiling-mediated repression gates downstream activation — anshulkundaje · 2026-09-13
- OpenAI open problem #2 cracked by human theorists; paper unifies and improves MRRW bounds — abeirami · 2026-09-13
- Redditor claims GPT-6 solved 3D two-phase conductivity-function closure, unverified — Severe-Ad8673 · 2026-09-13
- Dead fly's brain booted to read letters: 166,700 neurons simulated live — alvelda · 2026-09-13
- Patched vLLM+FlashInfer Pushes Gemma 4 31B to 150 tok/s on a Single B300, Beating SGLang — abhijithneil · 2026-09-13