If you’ve spent money trying to speed up your website and the scores barely budged, you’re not alone — and it may not be your fault. Google Search Advocate John Mueller recently spotlighted a case study explaining why so many Largest Contentful Paint (LCP) fixes miss the mark: teams optimize the wrong element entirely. It’s a technical-sounding problem with a very practical lesson for any business owner paying for site performance.
What happened
LCP — Largest Contentful Paint — is one of Google’s Core Web Vitals, measuring how fast the biggest visible element on a page loads. It’s a real ranking and user-experience factor, which is why businesses invest in improving it. According to
Search Engine Journal, Mueller pointed to a Google case study (published June 24 on web.dev) documenting a year of Core Web Vitals work at the e-commerce platform Nuvemshop. The surprising culprit wasn’t heavy images or slow servers — it was
which element the browser was measuring. The specifics:
- CSS transitions on carousels and banners delayed when elements became visible to the browser’s LCP detection — an issue present on 85% of storefronts.
- Because of that, browsers often locked onto the wrong element as the LCP, so every optimization afterward targeted something that was never the real problem.
Once the team fixed the measurement issue, the results were dramatic. Over January 2025 to January 2026, Nuvemshop’s
“good” LCP scores rose from 57% to 96%, and its overall Core Web Vitals pass rate climbed from
48% to 72%. They also reported an
8.9% increase in mobile conversion rate for Brazilian stores from organic search. The three practical fixes the case study describes:
- Remove CSS transitions from the top of the page so content displays instantly.
- Remove
loading="lazy" from the first image in top sections. - Add
fetchpriority="high" to the likely LCP image, with logic to confirm it’s the right one.
Why it matters
This is a cautionary tale about paying for the wrong fix. A business can spend real money compressing images and upgrading hosting, watch the score stay stuck, and conclude that “site speed is a black hole.” The Nuvemshop case shows the actual issue was often diagnostic: nobody confirmed which element Google was even measuring. Get the measurement right, and the gains compound — not just in scores, but in conversions, as Nuvemshop’s mobile numbers show.
After correcting which element the browser measured, Nuvemshop’s “good” LCP scores rose from 57% to 96% — with an 8.9% lift in mobile conversion rate. — via Search Engine Journal
What this means for small businesses
You don’t need to write code to act on this — you need to ask better questions of whoever manages your site:
- Confirm what’s being measured. Before optimizing, verify which element Google counts as your LCP. Fixing the wrong element is wasted money.
- Suspect your sliders and banners. Fancy homepage carousels and fade-in animations (CSS transitions) are a common hidden cause of poor LCP.
- Tie speed to sales, not just scores. The real payoff isn’t a green checkmark — it’s the conversion lift that came with it. Frame the work around revenue.
- Ask for before-and-after proof. A good web partner should show measured improvement, not just claim a page “feels faster.”
For a small business, a faster site means more visitors who stay, more calls, and more sales — but only if the fixes target the right thing.
The bottom line
Google just handed businesses a valuable diagnosis: many site-speed efforts fail because they optimize the wrong element, not because speed can’t be improved. Make sure your team confirms what’s actually being measured before spending a dollar on performance — because when the fix is aimed correctly, the results, as Nuvemshop found, show up in both your scores and your sales.