Guides
What actually happens between keypress and pixel
Five stages, one honest read on your rendering pipeline. Here is what each stage costs, what usually stalls it, and the single change that moves your numbers.
Tag
core web vitals
LCP, INP and CLS: what each one actually measures, which of them your users notice, and how far a single fix moves them.
16 posts1 authorLast post 2 days ago
Guides
Preloading fonts without making everything else wait
A preload tag moves a font to the front of the queue. It also moves everything behind it further back, which is the part nobody mentions.
Guides
Bundle math: how much JavaScript you actually need
The 170 KB budget is a number someone tweeted in 2016. Your real number depends on your audience’s devices, their network, and how much of the page arrives as HTML.
Guides
Reading your own analytics before you optimise anything
Half the performance work teams do is aimed at a device nobody in their audience owns.
Performance
Why your site crawls on hotel wifi, and the four fixes that work
Latency ruins a page long before bandwidth does. Ship these four changes and most sites feel instant again by the next deploy.
Guides
How to run a performance review that changes something
A meeting where everyone agrees the site is slow is not a review. It is a shared feeling.
CSS
content-visibility, measured on a real page
One property that tells the browser to skip work for anything off screen. The numbers are good, and the caveats are real.
Performance
INP is not FID with a new name
FID measured the first interaction. INP measures the worst one, and that is a completely different bar to clear.