Magento has the highest performance ceiling of any platform we measured. The catch is that much of that speed ships switched off. Here is why, and what the data shows.
If your Magento store is slow, you have probably been told that Magento is just heavy, or that you need better hosting. Both can be true. The more useful explanation, and the one the data supports, is that Magento ships as a framework with its performance features turned off, and a fresh store runs on defaults that were never meant to be the finished state.
Magento isn’t slow. Unconfigured Magento is.
Start with the part that gets lost in Magento’s reputation. In our audit of 80,000 stores, the highest-performing stores in the entire benchmark ran on Magento. Its best stores reached a 90th-percentile score of 93.5, higher than any other platform’s ceiling, the hosted ones included.
The same platform also produced some of the worst. Magento had the widest spread of any platform in the study, a 10th-percentile score of 51.2 against that 90th-percentile of 93.5, and the largest share of stores in the benchmark’s lowest maturity tier, around 22%, nearly one in four. (The 2026 platform benchmark covers why that spread is the defining feature of a framework like Magento.)
A platform that produces both the fastest and the slowest stores is telling you something simple: the speed is available, and the default state does not switch it on.
Why Magento 2’s default settings are slow
Magento is not sold as a finished store. It is a framework built to be configured by a technical team, and that choice runs straight through its performance. The features that make Magento fast exist in the box, but they ship off, generic, or unset on a fresh deployment: full-page caching and Varnish, which serve a prebuilt version of a page instead of rebuilding it on every visit, the production deployment mode that precompiles and minifies static assets, CSS and JavaScript bundling, image optimization, browser caching, a CDN. A framework optimized for flexibility out of the box is, almost by definition, not optimized for speed out of the box.
That trade is reasonable on Magento’s terms. The problem is what happens to it in the wild.
What the benchmark shows
The data shows how many stores are still running in that underconfigured state. About half of the Magento stores we scanned, 9,974 of them, failed a basic browser-caching check, the kind of setting that should be standard on any production store, among the cheapest speed wins there is, and a clear marker of a store still running on partly configured defaults. Across the other speed signals the pattern repeats: the levers are available, and on a large share of stores they were never pulled.
Why it stays that way
None of this is visible from where a store owner sits. You do not see “full-page cache: off” on your storefront. You see a slow site and a list of plausible-sounding theories about hosting and themes. The levers usually take a developer to find and enable, so on a store without ongoing technical ownership the slow defaults simply persist, quietly, for years. Magento’s weight makes it less forgiving than a lighter platform, so the unconfigured state costs more here than it would elsewhere.
How to find what’s slowing your Magento store
The fix is not “switch platforms,” it is to find which of Magento’s performance features are still off. In practice that means checking:
- whether full-page caching and the production deployment mode are actually enabled (a question for whoever manages your install);
- whether browser caching is set on static assets;
- whether CSS and JavaScript are bundled, minified, and not render-blocking;
- whether images load lazily;
- whether server response time points at hosting or at configuration.
Why each of these matters for conversion and ranking is in the site speed guide; the part specific to Magento is that the platform hands you the levers and leaves them off.
The same checks that flagged those 9,974 browser-caching failures across the benchmark are the ones a scan runs on a single store, and they are the fastest way to see which of these your own store is still failing, the visible edge of the levers still switched off.