My Wordpress Pagespeed Optimization Guide

[Septabyte]

https://docs.google.com/document/d/1ncQcxnD-CxDk4h01QYyrlOh1lEYDS-DV/edit?usp=sharing&ouid=114514252262811175804&rtpof=true&sd=true

I've spent the last 5 years finding the content and compiling it into my guide. The gdoc is 148 pages. I'm certain that it is the most comprehensive guide on the internet. I've scoured every wordpress pagespeed optimization guide and many, many non-wordpress specific resources(server/web stack level optimizations) with a focus on including little known unconventional optimization strategies.

This is my personally built site:

url: https://strongbpc.com

https://imgur.com/a/ZHBrSe5

Website Speed Test Result: strongbpc.com

See how fast this page is and how it can be made faster.

www.debugbear.com

By following my guide and implementing every step/as many as you can, I'm positive that you can get almost any site you maintain to earn a score over 90+ on pagespeed insights.

If you'd like to discuss a consultation or hire me for pagespeed optimization services, please shoot me a dm to discuss. I'm also happy to tutor anyone who is interested in lessons on pagespeed optimization.

If you have any content addition requests for the guide, or you would like me to expand on something already in the guide, please don't hesitate to ask. I hope this is helpful to you guys!

Edit: The table of contents unfortunately does not open by default, click on this icon to open it.

Edit 2: Since it's been requested, an example of a more complex site I optimized is https://moderatelymessyrd.com. It's on low tier siteground hosting. I only spent a bit of time on this one so it could be slimmer still with more effort.

 

[Groot]

Good try!
Anyway, it's a website with a little bit content, no slideshow, no animation effects, I think you need a better case study to get hired, lol!
In real world, 80 -> 90 mobile points on GPI enough, not worth to take much time to get 99 or 100.

 

[Septabyte]

Good try!
Anyway, it's a website with a little bit content, no slideshow, no animation effects, I think you need a better case study to get hired, lol!
In real world, 80 -> 90 mobile points on GPI enough, not worth to take much time to get 99 or 100.

It's just my testing ground, I can do the same with a more complex site without too much difficulty. I've been working with what's there as a base to ensure everything is optimized properly, then adding content is simply an afterthought optimization wise. For my personal projects I build for speed first, content second. There's no rush to launch this site as an actual store, so content is currently not my priority.

I'm actually working on some other optimization jobs right now with more complex sites and I'll have links to showcase for those as well once the work is completed.

>In real world, 80 -> 90 mobile points on GPI enough, not worth to take much time to get 99 or 100.

I'm going to have to disagree with you there. 90+ is required to get over this hump:

Google: 53% of mobile users abandon sites that take over 3 seconds to load

The average load time for mobile websites on a 3G connection is 19 seconds.

marketingdive.com

Page Load Time in Relation to Bounce Rate

Discover the latest data on @ThinkwithGoogle

thinkwithgoogle.com

> The probability of a bounce increases 32% as page load time goes from 1 second to 3 seconds.

Pagespeed is critical to minimizing bounce rate and retaining traffic.

 

[Groot]

I've worked on web design business over ten years, at the first time, they want GPI point too high, ready pay extra money for that, we take much time and hard work to make it real and then websites running a time and more content added, GPI becomes lower and lower then they just forget it, because there are many other thing should do that are more important than speedtest scores.
Really, Page Speed is important, for SEO and UX, but 90 or 100 GPI just for fun, most popular websites don't care about it, it does not affect SEO as much as newbies think, even trying to optimize to achieve high GPI scores will have to reduce many features necessary for user experience.

 

[Septabyte]

I've worked on web design business over ten years, at the first time, they want GPI point too high, ready pay extra money for that, we take much time and hard work to make it real and then websites running a time and more content added, GPI becomes lower and lower then they just forget it, because there are many other thing should do that are more important than speedtest scores.
Really, Page Speed is important, for SEO and UX, but 90 or 100 GPI just for fun, most popular websites don't care about it, it does not affect SEO as much as newbies think, even trying to optimize to achieve high GPI scores will have to reduce many features necessary for user experience.

>it does not affect SEO as much as newbies think

It's not just SEO, it's retaining users once they're actually on the site, SEO is the entry point. Google will indeed rank the site better if it's faster. But further than that, once they click the link from Google and actually enter the site, pagespeed then helps with user retention. That's what the two articles I referenced explain, is that page load time affects bounce rates once the users are already on the site post-google.

 

[Themaster007]

You don't have a better step by step of how you configured your website to have a better guide.

 

[Septabyte]

You don't have a better step by step of how you configured your website to have a better guide.

If you mean how I configured the Elementor site I linked, the configuration is indeed in the guide already under Elementor in the page builders section.

 

[kalios]

Some useful resources in there Septabyte - thanks for sharing your hard work

 

[Septabyte]

Some useful resources in there Septabyte - thanks for sharing your hard work

Thanks man! Took me a looooooong time to put this together, so l appreciate the kinds words . Please let me know if you have any content addition requests, I'm always looking to improve it further and add more resources. I've got the need for speed.

 

[belgianangel]

Thanks for sharing this comprehensive resource! Really appreciate all your hard work!

 

[lachhab]

Good try!
Anyway, it's a website with a little bit content, no slideshow, no animation effects, I think you need a better case study to get hired, lol!
In real world, 80 -> 90 mobile points on GPI enough, not worth to take much time to get 99 or 100.

That's what bothers me as well, you'd get 99/100 in most cases with astra/kadence/gp/blocksy out of the box.
The website is minimal, no background images, no videos, no forms, no animations, no popups. It would be ideal if there was a more challenging case study.

Btw I appreciate the effort and I thank you for the guide!

 

[guguk]

I appreciated to share your experience and doc as well.

But unfortunatelly you are missing important part as always. You are focusing one way that "increasing speed" but you should firstly consider to "how to reduce things that decreasing of speed."

I realized that you havent use even image logo when I visit your website. And please allow me that I want to criticising with your website.

It's just my testing ground, I can do the same with a more complex site without too much difficulty. I've been working with what's there as a base to ensure everything is optimized properly, then adding content is simply an afterthought optimization wise. For my personal projects I build for speed first, content second. There's no rush to launch this site as an actual store, so content is currently not my priority.

I'm actually working on some other optimization jobs right now with more complex sites and I'll have links to showcase for those as well once the work is completed.

>In real world, 80 -> 90 mobile points on GPI enough, not worth to take much time to get 99 or 100.

I'm going to have to disagree with you there. 90+ is required to get over this hump:

Google: 53% of mobile users abandon sites that take over 3 seconds to load

The average load time for mobile websites on a 3G connection is 19 seconds.

marketingdive.com

Page Load Time in Relation to Bounce Rate

Discover the latest data on @ThinkwithGoogle

thinkwithgoogle.com

> The probability of a bounce increases 32% as page load time goes from 1 second to 3 seconds.

Pagespeed is critical to minimizing bounce rate and retaining traffic.

It's not true to required 90+ GPI. I can provide many "real" examples with my clients there have not even 40 GPI but they are making a mint of money

It's seem a landing page for selling a product. And there have 3 main functions:
- E-commerce functions (product page, cart and checkout page) --> most important one
- Blog page (archive and single post) --> Important for easily sell the product and gain visitors and potantial customers
- General Pages (static pages eg: about, contact) ---> Not important but should be there for more info your company

So, if you are focus ONLY speed you would missing most important parts like "cart and checkout page is working as usual or not". This website could be just example you have built it but the cart and checkout page does not work as usual. The ajax function does not working.

However, if you sell a product you need to give some extra info on single product page or landing page to your visitors like infographics, testimonials, video etc. So without those things you need to extra work also spend a lot of money on social media or another paid platforms. So, your works and times for spending optimization will be waste if you does not sell anything.
As 15+ years experienced web developer I can advise to you please start to optimization before creating a page. Just drop out to using "elementor" or something bloated builders. Focus hassle-free builders. People could not realized that but the funny part is I just realized that your font does not load when I visit your website even you are using "Jost" font.

As I mentioned on here: https://babia.to/threads/how-to-fix-speed-index-issue-on-wordpress-website.102870/post-1425924

I'm working as a web developer in a company and we need to provide many images, videos etc and my GPI is 98 WITHOUT breaking any part of site.

And I bet that you can't achieve same GPI with same content same design and same server environment with using elementor without using CDN.

 

[Septabyte]

I appreciated to share your experience and doc as well.

But unfortunatelly you are missing important part as always. You are focusing one way that "increasing speed" but you should firstly consider to "how to reduce things that decreasing of speed."

I realized that you havent use even image logo when I visit your website. And please allow me that I want to criticising with your website.


It's not true to required 90+ GPI. I can provide many "real" examples with my clients there have not even 40 GPI but they are making a mint of money

It's seem a landing page for selling a product. And there have 3 main functions:
- E-commerce functions (product page, cart and checkout page) --> most important one
- Blog page (archive and single post) --> Important for easily sell the product and gain visitors and potantial customers
- General Pages (static pages eg: about, contact) ---> Not important but should be there for more info your company

So, if you are focus ONLY speed you would missing most important parts like "cart and checkout page is working as usual or not". This website could be just example you have built it but the cart and checkout page does not work as usual. The ajax function does not working.

However, if you sell a product you need to give some extra info on single product page or landing page to your visitors like infographics, testimonials, video etc. So without those things you need to extra work also spend a lot of money on social media or another paid platforms. So, your works and times for spending optimization will be waste if you does not sell anything.
As 15+ years experienced web developer I can advise to you please start to optimization before creating a page. Just drop out to using "elementor" or something bloated builders. Focus hassle-free builders. People could not realized that but the funny part is I just realized that your font does not load when I visit your website even you are using "Jost" font.

As I mentioned on here: https://babia.to/threads/how-to-fix-speed-index-issue-on-wordpress-website.102870/post-1425924

I'm working as a web developer in a company and we need to provide many images, videos etc and my GPI is 98 WITHOUT breaking any part of site.

And I bet that you can't achieve same GPI with same content same design and same server environment with using elementor without using CDN.

The blank cart page with the blank textboxes you screenshotted was because I had previously had CheckoutWC installed and then removed it. I never got around to replacing it, I'm considering just going back to CheckoutWC. I intend to just use the Mini-Cart for the cart and there won't be a cart page listed, but I might make a basic one just as a backup in case someone somehow lands on the cart url.

>And I bet that you can't achieve same GPI with same content same design and same server environment with using elementor without using CDN

I don't have a CDN active on my site (and did not have one on when this test occurred), and personally, I actually avoid them. Please check the headers and urls in the debug bear report to confirm. You can get better pagespeed without a CDN than with one in the country the server is hosted in if the website is optimized properly and your users are nearby. If the site is small it generally isn't necessary if users are in the local region, but a site receiving a lot of traffic it's a must.

>However, if you sell a product you need to give some extra info on single product page or landing page to your visitors like infographics, testimonials, video etc. So without those things you need to extra work also spend a lot of money on social media or another paid platforms. So, your works and times for spending optimization will be waste if you does not sell anything.

All of the content is currently just placeholders. Those will be added when the site actually launches.

Edit: I added a Checkout and Cart page just for you . Optimized them as well.

 

[nulledking]

Example speed up website:

Doğtek Makine - Eşanjör Servisi, Boyler, Servis Yedek Parça

Eşanjör Servisi, Pompa, Boyler, Servis Yedek Parça konularında Doğtek Makine olarak hizmet sağlamaktayız. Tamir ve hizmetlerimiz için...

dogtek-phe.com

 

[Septabyte]

@nulledking

Solid job bro, high 90s. Which optimization plugins did you end up going with? I see WP Rocket for caching, but curious what you went with for the rest of your stack.

 

[nulledking]

@nulledking

Solid job bro, high 90s. Which optimization plugins did you end up going with? I see WP Rocket for caching, but curious what you went with for the rest of your stack.

I used wp rocket bro, this my portfolio

 

[Septabyte]

@nulledking

I saw WP Rocket, was wondering what else you used. Is it just WP Rocket or Perfmatters/Asset Cleanup etc as well? Looks like you're getting a lot of mileage out of Delay JS since I see none in your waterfall besides the lazyload js.

 

[Septabyte]

I used wp rocket bro, this my portfolio

You have taken care of the majority of the optimization opportunities on your site, the only things I can really recommend is that you subset your fonts to make them a bit smaller using a site like https://everythingfonts.com/subsetter or https://products.aspose.app/font/generator, compress your svgs using https://vecta.io, and compress your images with https://tinypng.com. If you do those, I think you're site could get to 100/100. You're in the very high 90s, compressing the images further alone could probably get you there. But the work to get to 100/100 you have left is minimal.

 

[cesareborgia]

@Septabyte Honestly, I appreciate the effort (and well, I laughed at the BPC tbh), but what I didn't get it (i just read some parts of it), was if you are actually using yourself all/almost all of this at the same time, since you're recommending people to do so.

If yes, then, the question you should be doing to yourself is why.
But then you mention that you avoid CDN's like the plague, when you start by preaching them.

A bit weird, but avoiding CDN's and saying they are a crunch is a bit, let's say, stupid?
Sorry, but there's no other word.

Because of this little thing:

WebPageTest

View this on WebPageTest.org...

www.webpagetest.org

+

WebPageTest

View this on WebPageTest.org...

www.webpagetest.org

Where is the speed?

Anyway, preloading the site is available on wp rocket, fpress and perfmatters, so makes no sense to say to do this twice.

As for the server, I wouldn't trust a host that is still running PHP 7.0.
Or a 10yr old CPU.
Or SATA SSD.
Or claiming that uplink when below the panel says otherwise (cheap version) or in the case of the "premium" when the hardware will not make use of it.

Ok, I stop.

As for @NullMaster, I think he's just trolling you (or maybe not), since he's just delaying the whole site to load, which is enough to troll Google too.
https://www.webpagetest.org/result/240426_AiDc14_2ZC/1/details

Edit. Just to clarify, i hit post by mistake, that's why the half of the post before and the other after.

 

[Septabyte]

@Septabyte Honestly, I appreciate the effort (and well, I laughed at the BPC tbh), but what I didn't get it (i just read some parts of it), was if you are actually using yourself all/almost all of this at the same time, since you're recommending people to do so.
If yes, then, the question you should be doing to yourself is why.
But then you mention that you avoid CDN's like the plague, when you start by preaching them.
A bit weird, but avoiding CDN's and saying they are a crunch is a bit, let's say, stupid?
Sorry, but there's no other word.
Because of this little thing:

WebPageTest

View this on WebPageTest.org...

www.webpagetest.org

+

WebPageTest

View this on WebPageTest.org...

www.webpagetest.org

>was if you are actually using yourself all/almost all of this at the same time, since you're recommending people to do so.

Yes I am! All of these techniques are complimentary, you should definitely implement as many/all that you can. My site is 174 kb (up from 168 in the screenshots since I reverted to gzip instead of brotli) because I've pruned out as much as possible. Pagespeed is a composite of many factors (all covered in my guide), all of them should be addressed.

>But then you mention that you avoid CDN's like the plague, when you start by preaching them.
A bit weird, but avoiding CDN's and saying they are a crunch is a bit, let's say, stupid?

I'll clarify, the majority of people are not going to implement the majority/all of the guide, and for those people a CDN will be beneficial. If your users are geographically spread out, then a CDN also may make sense to implement. I prefer to spend as little money as possible, and Cloudflare free makes my site slower. If I paid for their premium features, it would probably have a performance benefit or match the sites current performance. CDNs offer performance consistency across geographic regions since they have servers interspersed around major geographic regions. If I was going to serve traffic outside of the US I would need a CDN. I placed my server location in Dallas to mitigate some of the latency due to distance between users by putting it in a city in the central United States.

>Where is the speed?

Not sure what you're referring to in the pagespeed test, what are you referencing? The site loads in the low 1 second range on mobile for LCP, FCP and the Speed index.

>As for @NullMaster, I think he's just trolling you (or maybe not), since he's just delaying the whole site to load, which is enough to troll Google too.

Delaying javascript and unused css is a necessary optimization technique for fully optimizing a site. It isn't tricking Google, it makes the site load faster since the javascript and unused css is not downloaded until user interaction. This speeds up the initial rendering time and only downloads assets when they are needed.

>As for the server, I wouldn't trust a host that is still running PHP 7.0

VPSDime doesn't configure Wordpress or even preload their Linux VPS's with any PHP version, they give you the raw OS. My server is running PHP 8.3 configured via WordOPs.

>As for the server, I wouldn't trust a host that is still running PHP 7.0.
>Or a 10yr old CPU.
>Or SATA SSD.

The specs are bad on paper, but in practice it's fast hosting. I tested every cheaper and most mid range plans across every host on vpsbenchmarks.com, and VPSDime came out on top. This is through actual testing with my sites, not just benchmarks. Based on the benchmarks you would expect VPSDime to have middling performance, but it is superior to all the ones I've tested. For $7/month for a 4 vCORE 6 GB ram VPS it crushes all the competition assuming you aren't willing to break the bank for expensive hosting.

>Anyway, preloading the site is available on wp rocket, fpress and perfmatters, so makes no sense to say to do this twice.

What are you referring to?

>and well, I laughed at the BPC tbh

BPC-157 benefits, dosage, and side effects

BPC-157: The latest and most important research. Expert analysis and supporting evidence for practical effects, potential risks, and more.

examine.com

BPC-157 is amazing stuff if you haven't heard of it/tried it. I have fibromyalgia and it's basically the only reason I'm able to function.

edit: I see what you're referring to now on the second speed test. You set the testing location to Mumbai and the site is based in the central United States. I don't use a CDN, the page speed metrics are going to be worse due to the testing location you chose (which is what I was referring to above). Use Pagespeed Insights, Debug Bear or GTMetrix.