Data report · Published October 31, 2026 · Data from September 2026

State of WooCommerce Store Speed 2026

How fast do online stores load for real shoppers on phones in the US? We looked up Google's real-visitor speed data for 186,634 stores with a US audience and sorted them by the platform they run on. A typical WooCommerce store shows its main content in 2.2 s. A typical Shopify store does it in 1.3 s.

Key findings

1. WooCommerce stores are slower on phones than stores on hosted platforms. At the 75th percentile of real visits, the median WooCommerce store shows its main content in 2.2 s, against 1.3 s on Shopify and 1.6 s on BigCommerce. 62.4% of WooCommerce stores meet Google's 2.5-second mark, close to Magento (62.8%) and PrestaShop (61.9%), far behind Shopify (93.2%) and BigCommerce (94.6%). n = 53,437 WooCommerce, 116,358 Shopify stores.
2. Most of the gap is the server. The median WooCommerce store takes 1.4 s before the first byte of the page reaches a phone (75th percentile), Shopify 0.5 s. Only 22% of WooCommerce stores get under Google's 0.8-second guideline. Self-hosted Magento (18%) and PrestaShop (15.2%) look the same.
3. Stores with more plugins and more JavaScript are slower. 72.6% of WooCommerce stores with 0 to 2 detectable plugins pass the 2.5-second mark; with 10 to 14 it is 43.3%. Stores sending phones under 500 KB of JavaScript pass 72% of the time, stores sending over 2 MB 48.7%. n = 1,013 stores with 10–14 plugins; 6,647 with 2 MB+ of JavaScript.
4. A lot of store software is old. Of the WooCommerce stores whose version we could read, 23% are one major version behind and 9.3% are two or more behind. Of the 7,784 that announce their PHP version, 29.6% run PHP 8.1 or older, which gets no security fixes. Among 2,817 PrestaShop stores we checked ourselves, 60.2% run 1.6, 1.7 or 8.0–8.1, none of which are patched any more.
5. Where a store is hosted lines up with how fast it is. 72% of WooCommerce stores on managed WordPress hosts pass the 2.5-second mark, against 62.9% on mass-market web hosts and 56.2% where no host could be identified. n = 12,958 / 17,103 / 20,040.

None of this proves cause. A store with 12 plugins is often a bigger store with more going on, and stores on managed hosting tend to have someone looking after them. The method section lists what the data can and cannot tell you.

How fast stores load, by platform

The measure is Largest Contentful Paint (LCP): the time until the biggest thing on the screen, usually the main photo or headline, has appeared. Google's Chrome browser records it on real visits, and Google calls 2.5 seconds or less "good". For each store we take the time that 75% of its phone visits from the US in September 2026 came in under. That is how Google judges a site.

Speed on phones by platform, US visitors, September 2026 (medians across stores of each store's 75th-percentile value)

Responsiveness to taps (INP) is not the problem for any platform: the median store on every platform sits at 75–100 ms, well inside Google's 200 ms. The difference is in how long the page takes to show up.

Is this just big brands on Shopify?

Partly a fair question: Shopify's list includes many large, well-funded brands. Chrome groups sites into popularity bands, so we compared stores inside the same band. WooCommerce trails Shopify in every band, from the top 50,000 sites down to the long tail.

Share of stores with main content within 2.5 s, inside the same Chrome popularity band

The server is where WooCommerce loses time

Time to First Byte (TTFB) is how long a phone waits before the first piece of the page arrives: the connection, then the server building the page. On Shopify and BigCommerce that work is done by the platform's own servers and caches. On WooCommerce, Magento and PrestaShop it depends on the store's hosting, its caching and how much the site has to do for every visit.

A slow first byte pushes everything else back. When the median WooCommerce store's first byte takes 1.4 s, the page cannot show its main photo at 1.3 s like the median Shopify store, however light the theme is. In our own work this is usually missing page caching, an overloaded shared server, or plugins that run heavy database queries on every page.

Hosting type

The crawler that HTTP Archive uses recognises some hosting companies from the page's headers. We grouped them into types; the list of names in each group is in the method section. Hosting was identified for 34,085 of 54,573 WooCommerce stores.

WooCommerce speed by hosting type

A content delivery network on its own does less than people expect: 64.1% of WooCommerce stores behind Cloudflare pass, against 59.2% with no CDN detected. A CDN serves photos and files from nearby, but most setups still send every page request back to the store's own server.

Plugins, JavaScript and page weight

HTTP Archive recognises a few hundred well-known WordPress plugins from what they put on the home page, so the count below is "plugins we can see from outside", not the true number installed. The median WooCommerce store shows 5. Real stores usually run many more.

WooCommerce speed by total home page weight

The 15-plus group (36 stores) is left out of the chart because it is small; it sits at 41.7%, in line with the 10–14 group. Every step up in plugins or JavaScript costs a few points. No single plugin is the culprit in most stores we open; it is the pile.

Old software: how common it is

WooCommerce

WooCommerce puts out about one major version a year and does not publish end-of-life dates, so we do not call any version "end of life". We count how far behind the current major (11.x) each store is. HTTP Archive could read the version on 43,419 of 54,573 stores.

Being behind on WooCommerce barely changes speed in this data: 58.9% of stores two or more majors behind pass the 2.5-second mark, against 62.8% on the current major. Old WooCommerce is mostly a security and compatibility problem. The speed comes from the server, the theme and the plugins.

PHP

PHP is the language WordPress, WooCommerce, Magento and PrestaShop run on. Per php.net, PHP 8.2 and newer still get security fixes; 8.1 and older no longer do. Only 7,784 of 54,573 WooCommerce stores (14.3%) tell visitors their PHP version, so this is a sample of stores that leave that header on, which is often a sign of a server nobody has tidied in a while.

These stores are also a little slower: 49.9% of those on PHP 8.1 or older pass the 2.5-second mark, against 55.3% on a supported PHP. Newer PHP is faster at the same work, but an old PHP version usually comes with an old server and an old setup, so we would not credit the whole gap to PHP.

WordPress

On the 28,791 WooCommerce stores that show it, 61.8% run the current 7.1 branch, 19.6% run 7.0 and 18.7% run something older.

PrestaShop and Magento

HTTP Archive does not read PrestaShop versions, so we used our own crawler. It visited the home pages of 2,940 working PrestaShop stores with a US audience between September 6 and October 23, 2026, and read the version from the page's code on 2,817 of them.

60.2% of these stores run a PrestaShop version that no longer gets fixes. That is a floor: most "8 or newer" stores do not show whether they run 8.0, 8.1 (both unpatched) or 8.2 and up. On Magento, 80 of the 3,956 stores whose major version HTTP Archive could read still run Magento 1 (2%), which Adobe no longer supports.

What the seconds are worth

We did not measure sales, so the figures here are someone else's. In "Milliseconds Make Millions", a study Deloitte ran for Google in 2020, the mobile sites of 37 European and American brands were followed over 30 million sessions. When a site got 0.1 seconds faster across four speed measures, retail sites saw conversions rise 8.4% and order values rise 9.2% on average. Those are averages for large brands, not a promise for a small store, but the direction is the same everywhere: shoppers on phones give up on slow pages.

If your store is in the slower part of these charts, the cheapest first step is to measure it. Our free store speed test runs a home, category, product and cart page on a simulated phone and tells you, in plain words, where the time goes. The version checker tells you in a few seconds whether your WooCommerce, WordPress, PrestaShop or PHP is behind.

Test my store's speed Is my store outdated?

Method

Speed data. Google's Chrome UX Report (CrUX), monthly country table for September 2026 (chrome-ux-report.materialized.country_summary, yyyymm = 202609), United States, phones. These are real Chrome users who opted in to sharing usage statistics, not lab tests. Each value is a store's 75th percentile across all its page views that month, not just the home page. CrUX reports LCP, FCP and TTFB in 100 ms steps. "Good" thresholds are Google's: LCP 2.5 s, INP 200 ms, CLS 0.1, TTFB 0.8 s.

Which stores. The HTTP Archive crawl of September 1, 2026, phone client, home pages only, where its detector (Wappalyzer rules) found WooCommerce, Shopify, BigCommerce, Magento or PrestaShop. We kept the sites that also appear in the US phone CrUX table for September. That gave 186,921 rows. We dropped 131 sites that matched more than one platform and 156 duplicates where the www and bare address were both listed, leaving 186,634 stores. Stores with no LCP value that month count in the store totals but not in the LCP figures.

"US audience", not "US business". CrUX lists a site for the US when enough Chrome users in the US visited it. Some of these stores are based elsewhere and sell to Americans. This is the slice of stores that both datasets cover, not a count of all US stores.

Are they real stores? A site counts as WooCommerce when the plugin's code is on the home page, which also catches some donation pages, membership sites and shops that have closed. Our own crawler had visited 8,633 of these WooCommerce sites in September and October 2026; 92.1% showed a working cart. In that group the median LCP was 2.3 s for confirmed shops and 2.3 s for the rest, so the non-shops do not move the result.

Versions. WooCommerce, WordPress, PHP and Magento versions are HTTP Archive's reading of the home page. We checked its WooCommerce reading against our own crawler on 6,962 stores: the major version agreed on 91%, and in most disagreements (490 of 627) our later visit found a newer version, so the store had most likely updated in between. We counted versions 1.x, 2.x and anything above 11 as unreadable, since those are almost always another add-on's version number; they are never counted as outdated. PHP versions come only from stores that announce them. PrestaShop versions come from our crawler, which reads them from the page's code; it can tell 1.6 from 1.7 from "8 or newer" and sometimes the exact 8.x/9.x line. Support status: php.net, PrestaShop's release cycle, wordpress.org, all read on October 31, 2026.

Hosting, CDN and plugins are HTTP Archive's detections. Managed WordPress hosts: Cloudways, Convesio, Elementor Cloud, Flywheel, GridPane, Kinsta, Liquid Web, Nexcess, Pagely, Pantheon, Pressable, Pressidium, Rocket.net, SpinupWP, WP Engine, WordPress VIP, WordPress.com, wp.cloud. Mass-market web hosts: A2 Hosting, Bluehost, DreamHost, FastComet, GoDaddy, GreenGeeks, HostPapa, Hostgator, Hostinger, Hostwinds, IONOS, InMotion Hosting, Namecheap, Network Solutions, One.com, SiteGround, Strato, Web.com, iPage. Cloud infrastructure: Amazon Web Services, Azure, DigitalOcean Spaces, Google Cloud, Hetzner, Linode, OVHcloud, Oracle Cloud, Vultr. A store matched in more than one group goes to the first in that order. Cloud detections sometimes come from where a site keeps its files rather than where it runs. Page weight and JavaScript come from HTTP Archive's single phone test of the home page.

Statistics. Medians across stores, never averages. We show no group with fewer than 5 stores, and no chart bar under 30. No store is named anywhere in this report or the files. Everything here is correlation: the data shows which stores are slower, not why.

Reproduce it. The BigQuery SQL (about 30 GB scanned, roughly $0.20 at on-demand prices), the computed results (JSON) and every table as one CSV file. The CSV and JSON are free to reuse under CC BY 4.0 with a link to this page. CrUX and HTTP Archive data belong to their publishers.

SwiftForge Web (2026). State of WooCommerce Store Speed 2026. https://swiftforgeweb.com/reports/woocommerce-store-speed-2026/