See which plugins and methods increased site page load speed the most in this head-to-head test.
Table of Contents
- Caching vs Optimization
- About the Tests
- Plugins Tested
- About the Testers
- WebPage Test Results
- WebPage Test Desktop First View
- WebPage Test Desktop Repeat View
- WebPage Test Desktop Content Requests First View
- WebPage Test Desktop Content Requests Repeat View
- Google PageSpeed Insight Scores
- GPSI Lab Data Metrics
- Chrome Dev Tools Metrics
- Overall Top Performers Summary
- My Top Pick for Best Free Caching Plugin 2020
Caching vs Optimization
It’s super important to understand the difference in these two ways of speeding up your site load time.
Some of the plugins tested only have one or the other options, and some plugins have both.
You need to know which of them to use for which issue, and what order to do them to make your site fast.
READ: Caching vs Optimization for Site Speed for details.
About the Tests
I’ve been running deep case studies for several years now. And I have hundreds of site audit clients with all manner of variety in both themes and plugins. Plus, I work behind the scenes with plugin and theme developers, as well as hosts to help them make the best wares possible for us.
All that to say – I know my testing methods are top-notch and accurate because I have to deliver results for my clients. That includes accurately identifying what is slowing down the site and what will fix it.
If you would like to replicate my tests, you’ll find a brief on my baseline setup and testing methods listed below.
If you would like to see the full case study results and learn how to do deep testing and site audits like I do, you can!!! In my Webmaster Training you’ll learn how to fully secure a site and make it fast! You’ll also get access to the full case studies.
Testing Conditions
Below is a brief of the testing conditions I used.
Site Setup
- Hosting – NameHero (LiteSpeed server)
- Brunch Pro – Genesis theme
- Standard Post page with 4 large images
- Sidebar widgets
- Search
- Recent posts – text list
- Recent posts – 4 images
- Popular posts – 2 images
- Footer widgets
- Bio
- Recent posts – 3 images
- Plugins and widgets
- Social Sharing – Grow by Mediavine (Social Pug free) – inline content share
- Simple Social Icons – not just for Genesis themes
- MailMunch
Testing Process
The site was wiped clean of any orphaned directives, files, folders, or database tables that might have been left when each plugin was deleted.
The post URL was changed for each new test so that there would be no possibility of the online testers caching the post from the last test.
The tests were run over the course of a several weeks, so fluctuations in internet conditions may have been slightly different day to day.
Image of Test Post
Plugins Tested
Following are the free versions of plugins and code I tested:
- Baseline – refers to the results on my testing site without any caching or optimization
- WP Super Cache
- Hummingbird
- Breeze
- WP-Optimize
- Hard Code Cache – placed in the .htaccess file
- Autoptimize
- Autoptimize with Hard Code Cache
- Swift Performance Lite
- WP Fastest Cache
- LiteSpeed Cache
If you don’t see a popular plugin on the list, it was omitted on purpose. If you want to know why just ask – and see the comments below for those who have already asked about W3TC.
One additional note about Autoptimize. The developer, Frank Goossen was kind enough to contact me about the test results of AO when combined with hard coded cache. We also tested AO combined with WP Super Cache. The results were a bit mixed between the testers. I will be doing a deep dive into this matter, as I think AO most certainly serves a great purpose as a super optimizer for those who want to bring their own custom browser caching.
About the Testers
I used 3 testers:
- WebPage Test – for Desktop speed tests
- Google PageSpeed Insights – for Lighthouse testing on mobile
- Chrome Dev Tools – for real-world speed on my own ISP
Tester and Data Tutorials
See the following posts for a primer on how to run accurate tests and interpret the data.
- How to Run a Site Speed Test Accurately
- Online Speed Testers: How to Read the Metrics
- Google Core web Vitals: What you Need to Know and Do
WebPage Test Results
Following are brief explanations of how the plugins performed in each tester.
WebPage Test Desktop First View
First View tests check the optimization aspect of the plugins.
The following plugins have little first view optimization aspect. They are mostly caching:
- WP Super Cache
- Hard Code Cache
First Byte and Start Render are the most important, as they contribute directly to perceived speed.
Top 6 TTFB:
- WP Fastest Cache
- WP-Optimize
- Hummingbird
- LiteSpeedCache
- Breeze
Top 6 Start Render:
- Hummingbird
- Breeze
- WP Fastest Cache
- WP-Optimize
- Swift Performance Lite
- LiteSpeed Cache
WebPage Test Desktop Repeat View Results Chart
Repeat view is a second page view or a return visit to the site within a short time frame, usually less than one week.
All plugins/code should impact these results, as they tell the browser what to cache and for how long.
First Byte and Start Render are the most important, as they are perceived speed.
But the on_load time should be significantly lower than the first view time too.
Top 5 TTFB:
- Swift Performance Lite
- WP Fastest Cache
- WP Super Cache
- LiteSpeed Cache
- Breeze
Top 6 Start Render:
- Hummingbird
- WP Fastest Cache
- Breeze
- WP-Optimize
- LiteSpeed Cache
- Swift Performance Lite (was nearly tied with LiteSpeed)
WebPage Test Desktop Content Requests First View Results Chart
First view content requests on the baseline site are below and represent the data shown in this results chart.
Following is how each plugin/method performed.
The results chart indicates how each plugin reduced the number of requests on the first view.
The following plugins have little first view optimization aspect.
The following plugins are mostly caching:
- WP Super Cache
- Hard Code Cache
None of the plugins had any effect on fonts, as they are all Google fonts and come from Google and cannot be cached or improved.
Image reduction is usually the biggest improvement for speed. Most free plugins do not have an impact on images as they usually don’t contain lazy load. (Image file size optimization was not turned on with any plugin, even if it was available.) But a few of the free plugins do have lazy load and did improve load.
Free plugins with lazy load:
- Autoptimize
- Autoptimize with Hard Code Cache
- Swift Performance Lite
JS and CSS file optimization are the next most important factor.
Since CSS optimization is near equal, the following list is prioritized on JS optimization.
The top 4 plugins for 1st View Requests:
- Breeze
- Autoptimize
- Autoptimize with Hard Code Cache
- WP Fastest Cache
WebPage Test Desktop Content Requests Repeat View Results Chart
Repeat view content requests on the baseline site are below and represent the data shown in this results chart.
As you can see, there’s not much room for improvement, as GZip from the host took care of most of the caching directives.
The chart below shows the plugin results, with only one outlier, which was WP Fastest Cache. (I ran that test twice to be sure.)
Where you see blanks on the chart are 0 requests.
Following is how each plugin/method performed.
The outlier of Autoptimize with Hard Code threw off the scale so much that I had to use a logarithmic scale.
Where you see no data means the result was at 0.
Google PageSpeed Insights Scores
GPSI uses the Lighthouse tester.
And it tests on the mobile version of the site.
Not that we should pay attention to just the score, and instead pay attention to what makes them up, I knew you would want to see these results.
Also keep in mind that GPSI is a one-off tester and that a minimum of 3 tests have to be run and averaged for proper statistics. That was not done on these tests. So there is a margin of error between the plugin results, especially those that seem close.
Summary
Top 4 GPSI scores:
- Autoptimize
- WP Fastest Cache
- Autoptimize with Hard Code Cache
- Swift Performance Lite
GPSI Lab Data Metrics
The Lab Data on the baseline site are below and represent the data shown in this results chart.
The most important data points are:
- First Contentful Paint – when visitors start seeing anything
- First Meaningful Paint – when visitors start seeing the bulk of the content above the fold
- Time to Interactive – when the visitor can scroll
Following is how each plugin performed on this test.
Summary
Perceived speed is calculated in 1st Contentful Paint and 1st Meaningful Paint. These correspond to the WPT metrics of TTFB and Start Render, and are the most important.
Top 4 plugins for GSPI perceive speed:
- Autoptimize
- Autoptimize with Hard Code Cache
- Breeze
- Swift Performance Lite
The next most important metric is Time to Interactive, meaning how long a visitor has to wait to scroll.
Top 4 plugins for GPSI Time to Interactive:
- Swift Performance Lite
- Autoptimize
- WP Fastest Cache
- Autoptimize with Hard Code Cache
The Square result on the baseline site is below and represents the data shown in this results chart.
I selected the first square where the full site content could be seen.
Top 4 plugins for render:
- WP Super Cache
- Hummingbird
- Breeze
- WP-Optimize
Chrome Dev Tools Metrics
The following results were gleaned from the Network tab on using the Chrome browser in Dev Tools mode.
The baseline site results look like this:
These are the actual download times on my ISP, which is fiber optics and runs at 200Gbps speed.
That’s a fast time, so your results will vary based on your ISP speed.
Keep in mind this is the hard-line connection speed to my desktop computer. Wifi speeds from the router could be substantially slower, depending on your router.
Following is how each plugin performed in this test.
Summary
The load time at the end is the most important factor. It is to the on_load point of the browser when mobile is simulated.
This most closely equates to the GPSI load.
The Finish point is to the end of the document.
The top 2 plugins for Chrome load:
- Autoptimize
- Autoptimize with Hard Code Cache
The next top 3 plugins for Chrome load:
- Hummingbird
- WP Fastest Cache
- Litespeed Cache
Overall Top Performers Summary
The top 3 plugins from each category were given a points rank with:
- 3 points for coming in first
- 2 points for coming in second
- 1 point for coming in third
The graph below shows the overall winners and how many points they received in each category. (The taller the category color, the more points that plugin scored with it.)
The top 3 plugins are:
- Hummingbird – 12 points
- WP Fastest Cache – 10 points
- Autoptimize – 8 points
My Top Pick for Best Free Caching Plugin 2020
WP Fastest Cache.
It came in the top 3 over the most categories.
While Hummingbird scored best across the board, it also bloats your database like crazy. Over time, that will make your database slow to access. And that will make your page load slow.
While Autoptimize performed well, it was only tops across 3 categories tested, and that was in large part to it offering lazy load in its free version, which only impacts images. You will also notice that when browser caching was added to Autoptimize, it actually performed much worse. So, it’s impressive for the first view, but not for repeat/return views.
Why would you exclude W3 Total Cache from the list? It’s widely regarded as the best free caching plugin.
W3TC is a thing of beauty when it’s stable and properly configured. Too much instability over the years concerns me. And, after all these years and installations, it has never come out of beta. That should concern everyone. If it ever has a 1.x release, and is consistently updated, I’ll include it.
The only issue I’ve ever had with it is from the CSS/JS minification options. If you use it for browser caching, page caching and the object cache then you can use Autoptimize or Cloudflare for minification and it works flawlessly.
I’m glad to hear that for you. But when the community gets so hooked on a plugin that is not always maintained so that they have to split it off into a community supported version on GitHub, that is a great concern for me. Hope it does not become unstable for you in the future.
As mentioned, if it ever comes out of beta, I’ll reconsider. But a plugin that has stayed in beta for years is insane, IMO.
Thanks for this useful tutorial. Website optimization is a very important thing. I’m using the Autoptimize plugin that you mentioned in this post and I can surely say that this is one of the best plugins available for WordPress.
I really like Autoptimize too, and looking forward to doing tests on just it with deeper explanations of settings and combinations with caching. It’s certainly a good fit for more advanced site owners who want to customize the caching parts. And is one of the few free optimization plugins that has lazy load included. That’s a huge boon for speed!!
greaat content, thank you I needed this!