Why It Won't Be Found!
summary
Talk examines practical ways to improve search quality, focusing on evaluating and measuring search relevance beyond simplistic thumb-rule checks. It covers building an evaluation pipeline (a "valuator") for search quality, why off-the-shelf solutions aren’t available, and when to use human labeling platforms such as Yandex.Toloka or Amazon Mechanical Turk, with many concrete examples and minimal theory.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 3:17Why search fails
- 6:27Measuring search quality
- 19:16Analysis and metrics
- 27:49Specific improvements
- 40:28Long-term optimization
- 50:01Summary and Q&A
related talks
35mThings I Learned the Hard Way Building a Search Engine
47mNavigating the SEO Landscape: Trends and Tactics
37mSearch and Recommendations: 3 Sides of the Same Coin #bbuzz
47mThe Future of Search and SEO | Panel Discussion
42mImproving Search @scale with efficient query experimentation
07mSEO Black Holes • Lightning talk
41mSEO and the Future of Search
47mChoosing a Search Engine with Your Head