Ecommerce site speed: the lungs of your store
Ecommerce site speed depends on hosting, your platform, its plugins or apps and the weight of your photos. In the anatomy of your site, these are the lungs. Healthy lungs give your store the breath to go fast and far. Here’s what Google measures, what slows a store down, code bloat, the signs of fatigue and five checks you can do yourself. The SEO audit examines the lungs of your French store.
What Google measures: three thresholds for site speed
Google measures your visitors’ experience with three indicators: the Core Web Vitals. Its Search Central help page sets a threshold for each. LCP (Largest Contentful Paint) measures when the largest visible element appears: 2.5 seconds at most. INP (Interaction to Next Paint) measures the response to a click: under 200 milliseconds. CLS (Cumulative Layout Shift) measures elements that move during loading: a score under 0.1. A page passes when 75% of visits meet these thresholds. Mobile and desktop are measured separately.
For ranking, Google writes that its systems use the Core Web Vitals, without saying how much weight they carry. Its page experience documentation adds that a good score doesn’t guarantee a top spot. Google may prefer the most relevant page, even a slow one.
At Ahrefs, the maker of an SEO tool, technical SEO specialist Patrick Stox plays down their role in his article. In his view, they aren’t an important factor.
Hosting and WordPress: what slows a store down
According to the HTTP Archive’s 2025 Web Almanac, 48% of sites pass Google’s three thresholds on mobile.
Hosting comes first. Google’s developer site, web.dev, splits LCP into four phases. The first is the server response time. When it’s too long, reaching 2.5 seconds becomes hard, or even impossible.
WordPress and WooCommerce recommend PHP 8.3 or greater and HTTPS hosting. Below that, WooCommerce’s requirements warn of poor performance and possible security vulnerabilities. Outages count too. When your server returns errors, Google may slow down its visits to your site. If a page stays unreachable for a long time, it may drop out of its index.
Code bloat: from plugins to copy-pasted text
Code bloat is code the browser downloads without using it. It still has to be read and then run. It downloads at the same time as your photos and can delay their display. The Lighthouse documentation, Chrome’s analysis tool, describes it for JavaScript as well as CSS.
According to the HTTP Archive’s 2024 Web Almanac, 44% of a page’s JavaScript goes unused during loading (median on mobile).
On a WordPress store, this code comes mostly from plugins added over the years and from the theme. Lighthouse suggests reducing the number of plugins or replacing them. The WordPress documentation gives the same advice. On Shopify, the Shopify help center lists theme code, apps and images among the features that can slow a store down.
Your texts can contain bloat too. Copy-pasting from Word, Google Docs or ChatGPT leaves invisible tags: empty paragraphs, useless styles, attributes specific to ChatGPT. I removed many of them from texts published by Newp, a web agency I worked for. These tags weigh little. On November 6, 2026, I measured the French version of this page: the text made up 13.5 KB of 68 KB of HTML. Removing them mostly keeps the display clean and pages easy to edit.
Signs that your store’s lungs are getting tired
You may recognize yourself in one of these situations:
- Your product pages take a while to load on 4G.
- The big photo arrives after the text, in pieces.
- The page jumps just as you want to tap a button.
- The menu or the cart responds with a delay.
- A customer tells you the site was down.
- You’ve installed plugins or apps over the years and no longer know what some of them do.
Each of these signs may point to a problem with Google’s three thresholds, the hosting or unused code.
Five checks to do yourself, from PageSpeed Insights to Search Console
PageSpeed Insights is Google’s free tool for testing a page. It gives two kinds of results. Field data comes from real Chrome visitors, over the last 28 days. The lab test simulates a visit: it helps find the cause.
Here are the five checks:
- Open PageSpeed Insights and paste the address of a French product page, not just the homepage.
- Start with the mobile tab: Google assesses mobile and desktop performance separately.
- See whether data from real visitors appears, then whether the Core Web Vitals assessment passes.
- Spot which of the three indicators isn’t green. Further down, the lines “Reduce unused JavaScript” and “Reduce unused CSS” point to unused code.
- In Search Console, open the Core Web Vitals report: your pages are rated “Good”, “Need improvement” or “Poor”.
A recent or little-visited store sometimes doesn’t have enough visitors for field data. Only the simulated test then appears. The fix depends on your hosting, your theme and your plugins or apps: that’s the audit’s job.
Beauty and wellness brands: photos that weigh a lot
A cosmetic brand shows its shades, textures and bottles from several angles. For supplements, it’s the tubs, labels and flavors. When the big photo is the largest visible element on the product page, it’s the one LCP measures. On mobile, the largest element on a page is an image in 76% of cases (2025 Web Almanac). The same logic applies to the big image at the top of your homepage.
Your visuals sell: keep their quality and lighten the file instead. The web.dev guide to LCP gives three leads: an image at the right size for each screen, modern formats like AVIF or WebP, and compression. Google also accepts these formats in Google Images.
From the lungs to a full check-up
The SEO audit checks the speed, uptime and mobile display of your French store. You leave with three priority problems and a 30-day action plan. It also spots unused code. I can clean up the code in your texts myself. For plugins and apps, I tell you which ones to remove or replace.
The free check-up starts elsewhere: your French homepage, your French keywords, your product pages and your shoppers’ trust. Your result appears right away, no email needed.
The Doc’s references
| Idea | Source | Level | Date checked |
|---|---|---|---|
| Three indicators and their thresholds: LCP 2.5 seconds, INP 200 milliseconds, CLS 0.1 | Google Search Central, “Understanding Core Web Vitals and Google search results” · web.dev, “Web Vitals” (updated October 31, 2024) | 1 | November 17, 2026 |
| 75% of page loads, mobile and desktop measured separately | web.dev, “Web Vitals” (updated October 31, 2024) · Google, “About PageSpeed Insights” | 1 | November 17, 2026 |
| Used by ranking systems, with no guarantee; relevance can come first | Google Search Central, “Understanding Google page experience” | 1 | November 17, 2026 |
| Not an important ranking factor (expert opinion) | Patrick Stox, “Core Web Vitals”, Ahrefs (updated January 5, 2025) | 2 | November 17, 2026 |
| Image as the largest element on 76% of mobile pages; 48% of mobile sites pass the three thresholds | HTTP Archive, 2025 Web Almanac, Performance chapter (January 15, 2026) | 2 | November 17, 2026 |
| Server response time, the first phase of LCP; images at the right size, AVIF or WebP, compressed | web.dev, “Optimize Largest Contentful Paint” (updated March 31, 2025) | 1 | November 17, 2026 |
| Formats supported by Google Search, including WebP and AVIF | Google Search Central, “Image SEO best practices” | 1 | November 17, 2026 |
| PHP 8.3 or greater and HTTPS; otherwise, poor performance and possible vulnerabilities | WordPress.org, “Requirements” · WooCommerce, “Server Requirements” (WooCommerce 10.8) | 1 | November 17, 2026 |
| Server errors: Google slows down crawling; indexed URLs that keep failing are eventually dropped | Google, “How HTTP status codes affect Google’s crawlers” | 1 | November 17, 2026 |
| Field data over 28 days, lab data; no field data when there aren’t enough visits | Google, “About PageSpeed Insights” | 1 | November 17, 2026 |
| Search Console Core Web Vitals report: “Good”, “Need improvement”, “Poor” | Search Console Help, “Core Web Vitals report” | 1 | November 17, 2026 |
| Unused code (JavaScript and CSS) slows the page; reduce or replace the WordPress plugins that load it | Chrome, Lighthouse, “Reduce unused JavaScript” · Chrome, Lighthouse, “Reduce unused CSS” · WordPress, “Optimization” (updated October 14, 2026) | 1 | November 17, 2026 |
| Shopify: theme code, apps and images can slow a store down | Shopify Help Center, “Online store web performance” | 1 | November 17, 2026 |
| 44% of a page’s JavaScript unused during loading (median on mobile) | HTTP Archive, 2024 Web Almanac, JavaScript chapter | 2 | November 17, 2026 |
| Weight of the French version of this page: 13.5 KB of text out of 68 KB of HTML | Measurement of the public HTML of the French page, on November 6, 2026, before the code bloat section was added | Own observation | November 6, 2026 |
Level 1 means a primary source: an official text or the documentation of Google, WordPress, WooCommerce or Shopify. Level 2 means a reference recognized by the trade. An “own observation” is a check I did myself, with its date and conditions. My sources method explains how I choose and rank sources.