INP Replaces FID — What Changed
FID measured the delay before the browser could start handling the first input. INP measures the round-trip latency of every interaction, then reports the worst (98th percentile).T…
Read article →CLS Root Causes
The four classic CLS sources: images without explicit dimensions, ads and embeds, dynamically injected content above existing content, and fonts loading after the initial render.Fi…
Read article →AVIF vs WebP for Real Sites
AVIF typically delivers 30–50% smaller files than WebP at the same perceived quality, but encodes much more slowly.The right pattern: serve AVIF if the browser…
Read article →Font Subsetting Saves Real Bytes
A typical web font ships every Latin glyph plus extensive language coverage you do not use. Subsetting to just the characters you render trims it…
Read article →font-display: swap vs optional
font-display: swap shows fallback text immediately, swaps to the web font when ready. The cost is a visible swap on slow connections.font-display: optional gives the…
Read article →