Which plugin really makes your WordPress site faster, safer or more privacy-friendly? We measure it: on ordinary shared hosting in Germany, with Elementor, one setting at a time. You get numbers, settings you can copy and a clear recommendation.
Out of the box, WP Rocket was no faster in our lab than having no plugin at all. Properly configured it was, but a free combination got just as far. When the license still pays off, and what it costs.
WP Super Cache, Autoptimize and locally hosted Google Fonts matched WP Rocket in our test: 3.79 instead of 3.91 seconds LCP on mobile. Here’s how to set up the free stack, and what you give up.
Four cache plugins, one Elementor page, the same shared hosting. A page cache nearly halves server response time, but mobile load time barely moves. The big step came from loading Google Fonts locally, even with the free combination.
How to save your WP Rocket settings as a file and move them to another site. Plus our measured setup for Elementor as a template to download, tested in our lab.
With preloading on, the first visitor after a cache clear got our Elementor pages from the server in 392 instead of 884 milliseconds. It does take a while to finish, though, and some images ended up marked as “failed” in our list.
“Remove Unused CSS” is considered WP Rocket’s strongest CSS option. On our Elementor test page it saved 0.26 seconds on mobile and 20 requests, and the layout stayed intact. We also saw how it serves a page without any CSS at all.
Caching, images, Core Web Vitals: what makes your site measurably faster.
Make Elementor fast without giving up your design.
Backups that actually restore. Protection that doesn’t slow you down.
Cookie banners, fonts, analytics: GDPR questions answered with measurements.
SEO plugins compared, multilingual sites, structured data.
Forms without spam and emails that actually arrive.
Our free plugin shows you which plugins talk to outside servers, what they send, and who accesses your site from outside.
Every test starts on the same WordPress install, reset to a saved state first.
After every change we measure on mobile and desktop, five runs each. So you see what each setting really does.
We publish the median, with date and versions. The raw data is in the article.
Independent: no vendor pays for placements or results.