Asynchronous Personalization
summary
Claim: caching anonymous pages in the CDN and filling personalized bits via AJAX cuts perceived load and reduces backend load dramatically. Implementation uses Rails, CDN edge caching, a client-side asynchronous personalization layer, a tracking pixel to restore analytics, and a JS experiment library finch.js. Measured rough browser numbers show logged-in response dropping from ~3s to ~1.1s and anonymous initial response from ~1.8s to ~0.1s; the system works but forces more JS, new tests, and cookie/parameter plumbing.
AI-written summary · about
chapters (12)
- 0:00Intro
- 0:46overview
- 1:04performance anxiety
- 1:54super bowl party
- 5:06async personalization
- 7:45"Anonymous" user
- 9:15development implications
- 10:59onCurrentUserReady(callback)
- 12:38analytics
- 15:13a/b testing with Finch
- 16:59Finch.js
- 19:11the future
related talks
15mPersonalization, Performance, and Probably Dynamic Content
16mHigh-speed personalization with Next.js
39mApp performance
34mPredictive preloads for ultimate web performance?
58mCaching the uncacheable: delivering personalized experiences without sacrificing performance
47mCaching strategies on https://dev.to
25mCaching strategies on https://dev.to
26mFetching Data and the Perception of Speed