Google Search Advocate John Mueller highlighted a case study that exposes a critical flaw in how many websites approach Largest Contentful Paint (LCP) optimization. The study, published on web.dev on June 24, details a year-long effort by Brazilian e-commerce platform Nuvemshop. The result: the share of stores with good LCP scores jumped from 57% to 96%. But the real insight is not the improvement—it is the reason previous efforts failed.

The Hidden Problem: Browser LCP Element Selection

Nuvemshop’s team initially suspected image weight or server latency as the main issues. However, analysis revealed that the browser was selecting the wrong element as LCP. On Nuvemshop, merchants can organize homepage sections in any order, leading to carousels, banners, and product grids in varying positions. Carousels were present on 85% of storefronts. CSS transitions on carousels delayed their visibility, causing the browser to identify a banner further down the page as the LCP element. Consequently, all prior optimizations targeted elements that were never the LCP.