What page size should I aim for in 2026?
The HTTP Archive median web page is around 2.5MB, but performance leaders keep critical pages under 1MB. Mobile pages should target 500KB to 1MB for fast 4G load. Above 3MB the page becomes painful on slow connections, and bounce rate climbs sharply. Page size correlates strongly with LCP and FID Core Web Vitals scores. The fastest 10% of sites in any industry are typically half the size of their median competitors, which translates directly to higher conversions.
What contributes most to page weight?
Images dominate, typically 40-60% of total weight, followed by JavaScript at 20-30%, then fonts, CSS, and HTML. Video, when present, can swing this entirely. The pareto rule applies: optimizing the top 3 largest assets usually shrinks total page size by 50% or more. Run the page size checker, sort the asset list by bytes, and tackle hero images, third-party scripts, and large JS bundles first. Micro-optimizing 4KB CSS while a 2MB hero image lurks is wasted effort.
Why does my page size differ between page-size tools?
Tools differ in whether they count compressed (gzip/brotli) versus uncompressed bytes, whether they include third-party scripts, and whether they wait for lazy-loaded assets. Some report only what loads in the first 3 seconds, others wait for the full load event. Compare the same page in two tools and check their methodology before drawing conclusions. The most useful number is "bytes transferred over the wire" (compressed), since that is what users actually download.
How do I reduce page size without losing visual quality?
Convert images to WebP or AVIF (30-70% smaller than JPEG at the same quality), enable lazy loading for below-the-fold images, defer non-critical JavaScript, remove unused CSS (use PurgeCSS), serve text assets with brotli compression, and self-host fonts with WOFF2. Audit third-party scripts ruthlessly: every chat widget, ad pixel, and A/B test tag adds weight. Removing one unused third-party script often saves more than weeks of CSS optimization.
Does page size still matter when users have fiber internet?
Yes, for two reasons. First, mobile users on cellular networks are still 60%+ of traffic, and 4G or congested 5G can be slower than home wifi. Second, even on fiber, browsers must parse, compile, and execute JavaScript, which is CPU-bound, not bandwidth-bound. A 3MB JS bundle takes 2-5 seconds to execute on a mid-range Android phone regardless of connection speed. Page weight affects perceived performance in ways pure bandwidth cannot solve.
How often should I audit my page size?
Run a check monthly on key landing pages, and immediately after every code deploy. Page weight has a tendency to creep up silently as marketing teams add tracking pixels, A/B test scripts, and pop-up widgets without coordinating with engineering. Set a performance budget (for example, total page weight under 1.5MB, JavaScript under 300KB) and fail the build if a deploy exceeds it. Tools like Lighthouse CI or SpeedCurve enforce budgets automatically.