Empowering a Performance Culture
summary
Company consolidates 100+ sites into one app and cuts the homepage JS bundle from 360 KB to 27 KB, dropping load from ~13.5s to ~2.5s and improving First Contentful Paint by ~4s. Uses WebPageTest, Lighthouse and Perfume.js for metrics, swaps a JS modal for a CSS URL-based approach, defers code via intersection observer, changes clicks to touchstart shaving ~100ms, wrestles with 125 third‑party requests via tag manager, and ships a PR‑scanning tool called spotter to predict regressions.
AI-written summary · about
chapters (13)
- 0:00Introduction and background
- 2:10Executive buy-in for performance
- 4:19JavaScript bundle optimization
- 9:18Third-party scripts impact
- 11:50Developer awareness approach
- 15:45Optimizing user interactions
- 17:33Developer challenges and education
- 19:21Product and design collaboration
- 22:50Documenting trade-offs and debt
- 24:13Legacy code refactoring
- 25:39Predicting performance regressions
- 26:42Tools and developer experience
- 27:27Performance culture beyond code
related talks
48mThink Fast, First: Empowering a Performance Culture
25mGetting your Team Passionate About Web Performance to Achieve Performant Web Apps
33mPerformance culture. New opportunities with Embroider
16mDear non-performant app
37mSecret Shortcuts of Loading Web Performance
41mSpeed at Scale: Web Performance Tips and Tricks from the Trenches
28mPractical Performance
27mHow we make Web Apps lightning fast at Microsoft
