Deferring 3 scripts improved LCP by 0.9s When a script sits in the <head> without instructions, the browser stops building the page to download and run it. Think of it like a delivery driver blo…
Two preconnect lines shaved 300ms off LCP When your page pulls a font or script from another website, the browser must first 'introduce itself' to that server: look up the address, shake hands, agree …
Reserving ad space dropped CLS from 0.31 to 0.04 CLS — Cumulative Layout Shift — measures content jumping around while loading. The worst offender is usually an ad or banner that loads late and pushes…
Breaking up one script took INP from 510ms to 140ms INP — Interaction to Next Paint — is how quickly a page responds to a tap or click. The enemy is a 'long task': a chunk of work that hogs the browse…
Lazy-loading 40 images cut load weight by 2.6 MB Lazy loading means an image only downloads when you scroll near it. Think of it like a buffet that refills a tray only when you reach for it, instead o…
Why a 'green' lab score still failed real users There are two kinds of speed data. Lab data is a test in a controlled lab, on one device. Field data is what real visitors actually felt, gathered over …
