To test the difference in speed between WordPress and Jekyll, I set up two sites on an unimpressive shared hosting account. I added a large amount of text and a number of pictures to a post to have some data to work with. Once the pages were loaded with the same data I used GTmetrix and WebPagetest.org to test the sites.
The GTmetrix test of the WordPress site showed a page load time of 3.7 seconds. That’s not bad—better than most according to Google’s data. The GTmetrix test of the Jekyll generated site showed a page load time of 1.4 seconds. That’s 2.3 seconds faster or 164%. The PageSpeed Grade was 35% better and the YSlow grade was 8% better with the static site.
WebPagetest.org runs the site twice giving you a better idea of what a first time visitor would see and what a repeat visitor would see once your site has been cached. Here are the numbers:
Quite a bit different. The page load time is still faster with the static site—82% faster—but WebPagetest.org is showing a longer load time than GTmetrix.
For me, the decision to go with a static site generator is pretty obvious. The only question then is: do you need a database? The answer is “no” more often than you’d think, and if you can say no to the database I think you should. You’ll get more speed, more security, and less complexity, all of which should save you money.