Redefining Web Performance
summary
Argues that web performance is a cultural problem not just a metric and gives concrete evidence: median page weight 1.5 MB, 90th percentile 5,754 KB, JavaScript compressed median 352 KB (uncompressed ≈ 2,400 KB) causing ~1.2s CPU work median and up to 5.3s at the 90th percentile; cites tools and artifacts like YSlow/Firebug, Chrome trace, request map and WebPageTest; points out top‑down fixes (AMP, Apple News, Facebook Instant Articles) shift format but don’t fix root causes; recounts a million‑page CMS debt that required an expensive multi‑year overhaul.
AI-written summary · about
chapters (32)
- 0:00Intro
- 3:17Small Cocktail
- 4:45Tiny
- 5:18The Dictionary
- 5:45Bootylicious
- 6:00Literally vs figuratively
- 6:54Frontend performance
- 9:14Massive performance problem
- 9:29Network constrained
- 9:48GSMA
- 10:21Medians
- 11:02Devices
- 13:07Compressing JavaScript
- 13:22Doom
- 13:42CPU
- 14:09Web of Excess
- 15:15Performance is a Business Problem
- 15:40Request Map
- 16:48WordPress
- 17:21The Ark Times
- 19:11Performance and security
- 19:50Speculative execution
- 20:28Example of speculative execution
- 22:23Accessibility and performance
- 23:13The Kindle
- 24:41Accessibility Performance
- 25:14Keystroke Level Model
- 26:17Is Performance Really Just About Time
- 26:44Task Completion
- 27:37Performance Research
- 31:17Performance Problems
- 31:39Performance Problem
related talks
36mNobody cares about web performance. It's hard, expensive and unneeded
28mWhat we talk about when we say "web performances"
35mRevitalising the Web - The Importance of Web Performance
31mEvery Millisecond Counts: Performance & UX
53mThe Long-Tail of Performance
22mNew Perspectives on Performance
40m“Performance by Design”
37mMobile Web Performance