Headless CMS Performance: No Platform Beats Wix's 53% Pass Rate

"Go headless and your site will be lightning fast" is one of the most confident pitches in modern web development. The field data is more sober. The headless CMS sites we could measure pass Core Web Vitals modestly, not spectacularly, and every one of them trails the ordinary hosted website builders that headless is supposed to leave behind. Ghost passes on 50.8% of its mobile sites, Sanity on 47.1%, Prismic on 47.0%, and Contentful on 40.4%. None of them matches Wix or Squarespace, which pass on about 53%.

That gap matters because teams choose headless partly for speed, and speed is not abstract. A site that loads slowly loses visitors before they read a word, converts fewer of the ones who stay, and ranks lower in search, since Core Web Vitals are a confirmed Google ranking signal. If an architecture is picked for performance, it is fair to ask whether the performance actually shows up for real users.

So we checked it against the field. A headless CMS separates content from presentation: the CMS, whether Contentful, Sanity, Prismic or Strapi, serves content over an API, and a separate frontend renders it. We joined the CMS detected on each origin to how those origins perform for real Chrome users in Google's own field data. The result is muted in an instructive way.

The figures come from HTTP Archive's technology detection, which identifies the CMS on each origin, joined to Google's Chrome User Experience Report, which records how fast that origin is for the real people who visit it. This is field data from actual visits, not a single lab test. The study spans 23,747 origins across four headless platforms, with each per-CMS bucket running between roughly 3,000 and 9,000 sites, so the exact percentages are best read as directional and the durable signal is the pattern.

A site passes Core Web Vitals only when it is good on all three metrics together: Largest Contentful Paint under 2.5 seconds for loading, Interaction to Next Paint under 200 milliseconds for responsiveness, and Cumulative Layout Shift under 0.1 for visual stability. We measure at the 75th percentile on mobile, the standard Google uses, because it reflects the slower phones and networks a real audience brings rather than a developer's fast laptop.

Sanity, Prismic, and Ghost clear the roughly 42% web-wide average. Contentful sits right on it. None of them is a standout, and all of them trail the hosted website builders we measured separately, where Wix and Squarespace pass on about 53% of mobile sites. An API-first architecture, in the field, does not beat a well-tuned all-in-one builder.

For a team weighing the two approaches, that is a genuinely useful result. The all-in-one builder that owns its whole stack, from images to caching, is producing faster real-world experiences than the flexible, developer-favored headless setup. Headless earns its place on flexibility and a clean content API, but the speed premium many teams expect to come free with it is not showing up in the data.

The reason is structural. A headless CMS does not render anything. It just hands content to a frontend, and the frontend decides the speed. Most headless sites are built with a JavaScript framework like Next.js or Gatsby, and as our framework studies show, those frameworks carry their own responsiveness costs on mobile. Going headless does not buy performance by itself. It moves the performance question one layer over, to whatever framework you chose to render the content.

You can see the mechanism in the loading numbers. Contentful's relatively weak loading, 62.9% good LCP, is consistent with heavier JavaScript frontends that make the browser download, parse and hydrate a framework before the main content paints. Ghost scores best on loading at 88.1% good LCP precisely because it is not really headless: it ships a lightweight, mostly server-rendered theme out of the box, so there is far less client-side work between the request and the pixels. Loading, not responsiveness, is what separates these platforms, and it tracks how heavy each frontend is.

Ghost is included for contrast rather than as a pure headless CMS. It is a publishing platform with its own themed frontend, and its strong numbers come precisely from shipping a light, server-rendered theme, which is the opposite of the heavy-hydration approach that drags the others down. On desktop the whole picture lifts uniformly, with Contentful at 56.8% and Sanity at 63.5%, because faster hardware absorbs the JavaScript cost that a phone cannot. The mobile gap remains a frontend-JavaScript story, not a CMS one.

Ghost ships its own themed frontend rather than being a pure headless CMS.

If you are considering headless for speed, temper the expectation and aim the effort at the frontend, because that is where these numbers are set. The architecture gives you flexibility and a clean content API, and it can be genuinely fast, but only paired with a lightweight frontend: static generation and minimal client JavaScript. Bolt a heavy React single-page app onto a headless CMS and you land where Contentful's average site does, at a 40.4% pass rate and 62.9% good LCP, right on the web-wide norm. Keep the frontend light and you land where Ghost does, at 88.1% good LCP, precisely because it ships a mostly server-rendered theme.

And if the goal is simply a fast content site with minimal effort, the field data points somewhere plainer than headless. None of the four platforms here reaches the roughly 53% that Wix and Squarespace pass at, and a genuinely static generator such as Hugo or Jekyll, which in our static-site study posts around 90% good LCP, will likely beat a headless-plus-SPA stack with far less work. The lesson is not that headless is bad. It is that the frontend, not the CMS name on the invoice, decides whether Sanity's 47.1% or Ghost's 50.8% is what your visitors get.

Embed this research

Cite this study