Google Search Advocate John Mueller highlights a new issue that explains why some of the Biggest Content Paint’s improvements often fail to produce results. In a store-specific layout, the browser may focus on the wrong object, causing all subsequent configurations to point to something that was never an LCP in the first place.
The case study was published on web.dev on June 24, and describes a year of work by Core Web Vitals at the ecommerce site Nuvemshop. At first, the team blamed image weight or server latency as the main problems. However, their analysis points to how the browser has chosen the LCP feature. After improving element selection, image importance, lazy loading, and edge retention, Nuvemshop reports that the share of its stores with good LCP scores has increased from 57% to 96%.
How to Choose the Wrong Item
In Nuvemshop, sellers can arrange the sections of the homepage in any way they like, resulting in carousels, banners, and product grids appearing in different places in all different themes. According to the case study, carousels were present in 85% of stores, and this is where the measurement encountered a problem.
The problem is based on CSS changes in carousels and banners. These changes are delayed when features become visible in the browser’s LCP detection. When consumers saw the carousel loading first, the browser occasionally identified the banner at the bottom as an LCP feature because its visibility was not delayed to change. As a result, according to Nuvemshop, previous improvement efforts have focused on items that have never been LCP.
The investigation revealed three main causes. The transition was pushing detection to the wrong frame, lazy loading was delaying images over wraps that needed to load faster, and critical images were missing critical signals. Additionally, the ratings are from both categories and product pages, not just home pages, so the solution had to work across all different page types.
Three Amendments to the LCP
The case study covers the fourth change, the temporary storage of the limit to reduce the delay. However, the main feature detection fix involved three changes to how the top of the page renders and loads. Nuvemshop has used all three changes in all of its main themes and page types. These changes are straightforward, following standard web performance best practices.
The team rolled out CSS changes to the top sections to display instantly, allowing the browser to recognize them as LCP candidates without delay.
They pull again loading="lazy" except for the first image in those top categories. web.dev’s own guidance is vague on this, warning against loading an LCP image because it always adds a load delay.
The last episode was fetchpriority="high" to a potential LCP image, which tells the browser’s preload scanner to grab it soon. Nuvemshop claims to have wrapped that up in terms of authentication so the signal only fires when something is a potential LCP candidate. Google’s advice is to reserve one or two important image attributes, because if everything is in the foreground, nothing is.
What Else Nuvemshop Reports
Nuvemshop reports that its overall Core Web Vitals pass rate improved from 48% to 72% year-over-year, not just the LCP count.
The company has updated the same Brazilian stores that are active in both January 2025 and 2026. For mobile visitors arriving through organic Google searches, it reports an 8.9% increase in conversion rate and an 8.4% increase in cart engagement.
Keep in mind that these are self-reported statistics, and the purchase numbers are year-to-year comparisons for that group, not a controlled trial. Nuvemshop ties guidance to its speed performance by relying on Deloitte research commissioned by Google, which found that shaving 0.1 seconds off the loading time can increase store conversions by 8.4%.
Why This Matters
Before pressing another hero image, double check that you are upgrading the feature that the browser treats as LCP. This is especially important for template-driven or carousel-heavy builds, where it may not hold.
Our latest launch sits in the same spot. In Looking at Web Content Across CMS Platforms, HTTP Archive data showed real-world LCP tends to break when the platform is slow to let the browser find the main image, or when it forces that image to compete with other more important resources, not just when the page is big. Barry Pollard also went over how to track down an LCP issue in PageSpeed Insights before you touch any code.
The conversion link shows Nuvemshop’s description of the data. Core Web Vitals is a small ranking factor, and Google has emphasized relevance as the most important, so the quick store is better read as a buying advantage than a ranking lever.
Looking Forward
This is just one company’s self-reporting, so purchasing benefits are best read as guidelines. The process itself isn’t new either, as web.dev has been championing this kind of discovery and relevance work for years. Added to the case study is a worked example of identifying the actual LCP feature across layouts that vary from store to store, which is a difficult part of any customizable platform.
Featured Image: victoriatudos/Shutterstock
