Heree: LCP improves by only 10% despite 59% image size reduction? SSR hydration delay is the cause

·2026.11.20 09:01

Key point

A two-line fix resolving SSR hydration delay reduced LCP from 2.3 seconds to 0.48 seconds in real browser measurements.

1 / 15

Details

  • Despite a 59% reduction in image transfer size, LCP decreased by only 10%, indicating the bottleneck was Render Delay rather than download
  • The initial opacity value of the shared image component was included in the SSR HTML, causing images to remain transparent until hydration completed
  • A two-line fix using the disableLoader prop to set the initial opacity of the main image to 1 improved Render Delay by 89% (approximately 9.4x)
  • Hydration waiting status can be verified through Lighthouse segment analysis and server HTML validation via curl

Root Cause Analysis

  • Measurements using Lighthouse 12.8.2 CLI revealed a delay of over 12 seconds between image arrival and display on screen
  • The TTI point when React hydration ends coincided with the end of Render Delay, making LCP dependent on JS execution
  • The issue does not occur in CSR environments but arises in SSR because the initial state is fixed in the server HTML

Measurement and Verification

  • Discrepancies exist between Lighthouse simulation and real browser measurements; using the median of at least 5 runs is recommended
  • The Chrome address bar popup UI interferes with LCP measurement, so related flags need to be disabled
  • Comparing LCP segment ratios with TTI and checking the opacity attribute of server HTML via curl are key to diagnosis

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.