Subresource compression algorithms on React sites
97.7% of audited sites (9,353) serve Gzip-compressed subresources (79% good LCP), 91.1% serve Brotli (79% good LCP), 41.3% serve Zstandard (74% good LCP), and the 0.5% adopting Dictionary-Compressed Brotli (dcb) reach 100% good LCP (1.3s median).
91.1% of React sites use Brotli compression subresource HTTP compression, and 79% of those sites have a good LCP, against 79% of all React sites.
LCP pass rate
Sites with Brotli compression subresource HTTP compression have a good LCP most often: 79%, against 79% of all React sites. Sites with Zstandard (zstd) compression subresource compression have it least often, at 74%.
Subresource HTTP compression usage
Unencoded binary assets is the most used: 99.9% of React sites. Brotli compression is 91.1% of React sites.
LCP phase breakdown by subresource compression algorithm
Across the four LCP sub-parts in Chrome's field data, the largest gap between Brotli compression and Zstandard (zstd) compression sits in time to first byte (TTFB): 696 ms median for Brotli compression against 838 ms for Zstandard (zstd) compression. Overall median field LCP is 1.6s for Brotli compression (TTFB 696 ms, load delay 424 ms, load duration 152 ms, render delay 193 ms) versus 1.7s for Zstandard (zstd) compression.
Subresource HTTP compression per platform
Wix sets the subresource HTTP compression to Brotli compression on 100% of its sites, and 92% of Wix sites have a good LCP, against 79% of all sites. WordPress sets the subresource HTTP compression to unencoded binary assets on 100% of its sites, and 66% of WordPress sites have a good LCP.
Subresource compression or the platform
Does the platform explain the subresource HTTP compression difference?
On Wix, 92% of sites with Brotli compression subresource HTTP compression have a good LCP, against 92% of Wix sites with unencoded binary assets. Within Wix, sites with Brotli compression and unencoded binary assets have a good LCP about equally often. On WordPress, 64% of sites with Brotli compression subresource HTTP compression have a good LCP, against 66% of WordPress sites with unencoded binary assets. Within WordPress, sites with Brotli compression and unencoded binary assets have a good LCP about equally often.
Brotli, Zstandard (zstd), and Dictionary-Compressed Brotli (dcb) across page subresources
Nearly every page (99.9%) transfers pre-compressed binary formats like WebP, AVIF, and WOFF2 without extra Content-Encoding. For text subresources (JS, CSS, SVG, JSON), 97.7% of sites use Gzip (79% good LCP, 1.5s median), 91.1% use Brotli (79% good LCP), 41.3% include Zstandard responses (74% good LCP, often from Meta/Google third-party scripts), and 0.5% (47 sites, led by Shopify/Cloudflare) use shared-dictionary Brotli (dcb) to achieve 100% good LCP (1.3s median).
Understanding subresource compression algorithms
Enable Brotli (br) and Zstandard (zstd) at your CDN edge for all text-based subresources (text/css, application/javascript, image/svg+xml, application/json) to shrink transfer size by 15% to 25% over Gzip. Across all audited sites, 91.1% serve Brotli subresources (79% good LCP) and 0.5% serve Dictionary-Compressed Brotli (100% good LCP).
By Arjen Karel · 9,570 crawled React sites (1,979 with desktop Chrome field data, October 2026). How we measured.