We pulled the Lighthouse, RUM and revenue data of 22 Spanish online stores. The correlation between LCP and revenue is brutal. Here's the math, the diagnosis, and the 6 fixes that actually move the needle.
Tell us your idea and we'll send you a fixed quote in less than 24 hours. No commitment.
There's a number we keep showing prospective clients in our first call. We pulled it from an internal study across 22 Spanish e-commerce stores between €30k and €1.2M of monthly revenue. The number is this:
Every +1.0 second of LCP, on the same store, with the same traffic, costs roughly €4,200 of monthly revenue for a mid-size Spanish online shop.
That's not a stat from a SiteSpeed.io blog post about Walmart. That's the median we measured here, in Spain, in 2025, after controlling for ad spend, season and SKU mix.
This piece is the long version of why that's true, why nobody fixes it, and what to do about it on a Monday morning if you actually want the money back.
Largest Contentful Paint is the time it takes for the largest visible element on your page (usually the hero image or main product photo) to render. Google has been very public that the threshold is:
In our 22-store dataset, plotted against revenue per session, the curve looked roughly like:
| LCP (mobile) | Median conversion rate | Bounce rate |
|---|---|---|
| 1.5s | 3.4% | 31% |
| 2.5s | 2.6% | 39% |
| 3.5s | 1.9% | 47% |
| 4.5s | 1.4% | 54% |
| 5.5s+ | 0.9% | 63% |
For a store doing 30,000 mobile sessions/month at a €42 AOV, going from 3.5s → 1.5s is the difference between €23,940 and €42,840 in monthly revenue. That's the same store. Same products. Same campaigns. Just faster.
Performance is not a tech problem in Spain. It's a cultural one in 90% of the businesses we audit:
Meanwhile the site quietly accretes weight: a chatbot script, a heatmap, a popup tool, a YouTube embed, a TikTok pixel, two analytics pixels, a 6 MB hero video, a font with 9 weights nobody uses.
We've measured it: a "modern" Spanish e-commerce in 2025 ships an average of 5.4 MB of JavaScript on the homepage alone. The actual HTML is rarely over 80 KB.
We are going to skip the obvious advice you've read on every blog ("optimize images!"). Instead, here are the six things, ranked by impact-per-hour-of-work, that we apply to every Spanish e-commerce we touch.
This is the LCP killer. We see 6 MB JPEGs in 2026. The fix:
srcset at 480 / 768 / 1080 / 1440 widthsfetchpriority="high" on the hero <img><link rel="preload" as="image"> in the document head, only on the homepageA cosmetics brand we worked with: hero 4.1 MB → 280 KB AVIF. LCP from 3.8s → 1.6s. Done in a single afternoon. +18% conversion the next 30 days.
Heatmaps, chatbots, support widgets, even most analytics — none of them need to fire in the first 3 seconds. We:
<script async> and a 2-second delay (ConsentMode v2, of course)Single biggest impact-per-effort win after image optimization. Typically saves 800ms–1.4s of LCP.
A WooCommerce or Shopify theme bought from ThemeForest in 2022 is dragging in 4 page builders' worth of code. We see this every week.
You don't always need a full custom rebuild. Often a stripped-down theme purpose-built for your store, removing the 80% of features you don't use, is enough.
A Shopify store doing €120k/month migrated from a heavy theme to a custom Liquid build. Lighthouse 38 → 92. Revenue per session +24% the month after. No new traffic, no new ads. Just less code.
We routinely see e-commerces loading 7+ font weights when they only use Regular and Bold. Each weight is ~30–60 KB. Worse, they're often loaded as font-display: block, which forces the browser to wait before rendering text.
Fix:
font-weight: 600 if you must, but Bold + Regular covers 95% of UI.font-display: swap always<link rel="preload" as="font" type="font/woff2" crossorigin> on the one primary fontRoutinely saves 300–600 ms of LCP and a font-flash that looks cheap.
loading="lazy" is necessary, not sufficient. The right pattern is:
loading="lazy"preload="none" + a poster imageCloudflare's free plan is enough for most Spanish e-commerces under €1M/year. Make sure:
Cache-Control headers (with revalidation on stock changes — Shopify and WooCommerce both expose this)A Sevilla-based shop: TTFB 620ms → 110ms with Cloudflare turned on properly. LCP −400ms across the board.
Don't trust your dev's local Lighthouse run. They're on a 1Gbps fiber line in Vallecas. Your Madrid grandmother is on a slow 4G connection on the metro.
The real data:
Run all three. They will agree within 10%, and the truth will hurt.
Open your store on your phone. Cellular, not WiFi. Forget what you think it looks like — actually time it with a stopwatch from the moment you tap until the hero image is fully loaded.
If it's > 3 seconds, you're losing roughly the equivalent of one paid full-time employee in revenue per year, every year, for as long as it stays slow.
If you're a Spanish e-commerce doing €30k+/month and want a written, RUM-backed performance audit with the projected revenue impact of each fix, we'll do it free. You'll get:
Request a free performance audit or explore our web performance services.
Speed is money. We just measure exactly how much.