By: Tyler Bishop
Below, I'll address some of the most common questions and anecdotes about website speed in the context of Ezoic.
I'll also show you how to use Ezoic to see site speed EXACTLY how Google does from an SEO perspective using their preferred too, Google Lighthouse.
1.) My website "speed score" is...
Unfortunately, the most commonly used website speed tool used by publishers is Google's Pagespeed Insights — which has been broken since late 2018 and has yet to be fixed by Google's team even to this day. This means that the scores often show much lower than they are designed to show; which is by using Chrome Lighthouse inside the Developer Tools.
That said, what about GTMetrix, Lighthouse itself, and other pagespeed tools that are calculating a slow score on my site?
Before digging into how to troubleshoot and ensure that Ezoic is configured properly on your site, it is important to outline the limitations and dangers of relying on tools as a measure of actual website speed, or as an indicator of something that — if improved — would directly provide a benefit in some or fashion in terms of SEO, traffic, UX, and revenue.
Website speed is important, as we have all heard, but site scoring tools are simply proprietary measurements of a sites speed that include a general check of common issues or challenges. A fast site may actually get a better score than a slow site in many of these tools because of the way a site is built or developed.
Take CNN.com for example, head over to the site on a mobile device from a mobile network. I'm willing to bet it loaded instantly for you. However, when processed through common tools the site shows about as bad of a score as you can get.
"CNN.com built their mobile site for users, not tools. The site may load/process many items in the background, but for users, the experience is almost instant. This means tools see the site as slow but users see it as fast. The tools are not connected to anything ad or SEO related; however, the visitor's experience 100% is..."
Tools are built to give general feedback and general recommendations. That is why so much of the advice is irrelevant, wrong (will make some sites physically slower), or would break many sites if implemented as recommended. For example, removing some elements of the "render-blocking" CSS on one of my sites instantly ruins the theme on my homepage.
Prior to getting into what you can do about actual site speed using Ezoic, and how Ezoic improves actual site speed for visitors, it is vitally important to understand the way these tools work and why they are not good ultimate measures of website speed.
Tools can be a good way to get information about specific areas of your site that can be improved, but the Ezoic Site Speed app is actually designed to do all of these things automatically. However, even with all the recommendations in place, many tools still produce a low score... which is ultimately irrelevant if the site itself is fast and speed is not actually impacting UX.
The one recommendation we do make to publishers is to use the Ezoic Site Speed Measurement Tool to compare the "without Ezoic" version of your site to the Ezoic version.
In most cases, if your site is set up properly, the Ezoic version is actually objectively faster and more lightweight. If the site speed app is on and caching is enabled properly, there are few exceptions to this.
Note: The exception is if your original site has no ads on it. No site can escape the issue of ads impacting the overall total page load time; however, lazy-loading ads will ensure that the website remains just as fast to visitors (while showing a lower overall tools score and page load time)
Additionally, all ads slow down total site load time, but single-network solutions, like AdSense alone, won't affect sites as much as things included in Ezoic like Header Bidding, Google's Ad Exchange, and more. These are things that every major publisher in the world deals with and uses and there are no great improvements to be made.
This can be a good way of understanding that Ezoic is not actually impacting website speed. What's more, if you see a major discrepancy in the way this looks, it is possible something is not working properly on your configuration.
2.) How do I ensure Ezoic is set up to show my users the fastest version of my site?
Step 1: Ensure the site speed app is installed on your site from the Ezoic app store. More on setting it up here.
Step 2: Configure as many of the site speed elements that are available inside the app as a "Version" and then test the tool to ensure your site displays properly. You can try different versions to ensure that the Site Speed App doesn't cause issues with your site and activate and de-activate them depending on your results. Ultimately, we hope that all modules inside the app work great on your site. Once they do, activate the version that is working best.
NOTE: The site speed app does not instantly take effect. It may be up to 24 hours until all the implementations are live and running on the site (assuming caching is set up correctly, see below). Additionally, the site speed app uses smart technology to modify and change things over time to help improve speed for each user the more time the app is turned on.
Step 3: Ensure Caching is configured properly according to your rules and that it is ACTUALLY doing its job on your website. This starts by installing and configuring the caching app (more on that all here).
Ezoic's Caching App will allow users direct access to Ezoic's free CDN and site speed caching best practices.
"Ezoic will always obey the rules set on your site first. This means existing cache headers on your site may have pre-existing rules set to prevent Ezoic from caching a page or site with our CDN.
These rules could be set by a plugin or extension if you use a common CMS, like WordPress or Joomla. They could also be set at your host as well. This could cause your site to be slower because of these plugins or tools, not Ezoic."
To understand if this is happening or if caching is working on your site, look no further than Ezoic's Big Data Analytics. Simply click this link to view the dashboard below that will tell you if caching is working correctly.
If you are seeing a large number of "caching off" or "caching miss", that means that the Ezoic caching app is not doing its job. This is likely due to an existing cache header on your website.
But how do I fix that? Commonly, this is because of an existing CDN of plugin/service integrated with your website. A host/CDN or a plugin that affects the speed or the delivery (even website back-up tools) could be setting default values that force your pages to load from the source upon every visitor query. Ezoic can fix this but you'll have to set Ezoic to override existing rules, or fix the existing rules on your website according to how frequently different types of pages might update (most pages are static and can be cached for much longer than 24 hours).
Updating these settings can ensure that these elements work properly on your site as long as you're aware of the way this will work.
NOTE: Updating cache settings, etc. is not an instantaneous action. This means immediately running reports or monitoring analytics in real-time to see if it improves speed is not going to work. You need to wait up to 24 hours before monitoring the impact.
3.) What else can be done with site speed using Ezoic?
1: Integration method:
- Cloudflare or Nameservers integrations with Ezoic offer the fastest experience all-around
- WordPress Plugin integration is acceptable but will not be as fast as Nameservers or Cloudflare (to a very slight degree)
- Snippet integration is considerably slower and never recommended unless there is no alternative.
2: Check your mobile users for the actual impact
Few sites see any issues with desktop speed but notice most flags from tools or analytics coming from mobile devices on very slow connections.
Keep in mind 4G and 5G or wi-fi enabled mobile devices will often be faster than desktop. Notice that there is almost no difference between all these mobile users in engagement time and that the majority are on a very fast mobile connection anyway.
"In the example above — 50% of users are mobile — and of mobile users, 51% are on a very fast connection anyway."
Although mobile users represented in tools show very slow connections, few of your users may be experiencing this. A great way to check is by using Big Data's UX report on connections — click here to see it.
3.) See if any of this matters on your site anyway
Often, there is an obsession with speed tool scores when in reality users are getting the experience they want in 99% of cases.
Using Ezoic's Big Data Analytics, you can actually see what percentage of visitors are getting different load times and how that is impacting their experiences.
If users are seeing generally fast load times or times that aren't impacting their experience, it is very likely that traffic (SEO) and ad revenue are both being positively impacted by speed, not negatively impacted.
4.) You could try AMP
Often, regardless of what data says from tens of thousands of sites we directly work with and look at every day, publishers will be dead set on improving arbitrary speed scores in the hopes that it will improve SEO or something else.
Small — and even large — pagespeed tool score increases have not shown to have many benefits unless that correlate with the info listed in 1-3 in our experience. What's more, Ezoic publishers see 2-3x the traffic growth of the average publisher each year based on multiple internal studies we've conducted.
That said, if you want a very fast site, you can always convert it to AMP. Ezoic offers an AMP converter and AMP visits from Google search display instantly and are as fast as it gets.
Unfortunately, for many publishers that already have experiences with AMP, they know that the benefits they hoped would come from SEO don't really exist unless they're a Google News publisher. Additionally, AMP generally is a lower-earning mobile framework and publishers typically are very disappointed in the earnings.
However, if speed is more important than revenue and a publisher doesn't care if SEO is completely unaffected, AMP will definitely produce a faster website.
4.) How is Google SEO being impacted by my speed with Ezoic?
For this part, I'll share n offline conversation I had with a publisher that thought they might have a major site speed issue with Ezoic, only to learn that Ezoic had actually made their site speed perfect in the eyes of Google's SEO team DESPITE A BAD PAGESPEED INSIGHTS SCORE!
This publisher gave me permission to share this after they had reached out with SEO and site speed concerns. I have removed some of the content and information to bring forth the most important info from the conversation and protect his identity.
The takeaway from Google:
Optimizing for arbitrary numbers is not a data-driven approach (it is gambling). I think we’ve likely the data regarding Ezoic publishers who, on average (and as a whole), see major organic traffic increases due to the actual UX improvements our machines are able to help publishers achieve. Google wants the site to be fast for users. They measure this by looking at Chrome Lighthouse scores (not Pagespeed Insights) relative to users that visit your website. What's more, they aren't just looking at speed. They are looking at how speed is affecting UX for visitors with different connection speeds (i.e. do 4G visitors bounce 50% and 3G visitors bounce 90% of the time?)
Ezoic makes sites faster for actual users and the site speed app is designed to do this, not optimize for tools. This is not luck. Our engineers have spent a lot of time refining what speed aspects of websites actually impact SEO and experiences for visitors.
That being said, let’s dive deeper into what a data-driven approach looks like on your website.
1.) Your actual website speed and the impact on users and SEO when using Ezoic
Looking at pageview duration, your TTI has almost no impact on the visitor’s experience when we look at the actual data here.
The same goes for bounce rate. This is all in Big Data Analytics Under Site Speed.
Additionally, over 80% of your visitors are mobile. So, the concern has to be how speed is affecting those users who could be on a slow connection…
+75% of visitors are on a 4G or better connection. They are going to have no problem loading your site.
In fact, I just ran your site through GOOGLE LIGHTHOUSE on a simulated fast connection in a Chrome browser (76, the same as Googlebot now), and guess what the results were...
Congrats! You have a 100/100 pagespeed score according to Google!!!
You can get the same result:
- Open Chrome in Incognito mode (with no extensions)
- Click Inspect and open up Developer Tools - Lighthouse (Lighthouse will render Chrome extensions if you don’t disable them FYI)
- Click Audits and select Performance, Mobile
- Then, select No Throttling and run the audit to see how 4G visitors would see the site
That wasn’t just on mobile either. I get the same result on Desktop.
A lot of this is due to the optimizations made by Ezoic’s site speed app and our caching. In fact, when I run the desktop version through the non-Ezoic version of your site, here’s what I get…
[It produced a much lower score ---- 48/100 ---- Ezoic made this site twice as fast for actual visitors]
So, here’s the good news. We can use math and Google’s own words to highlight why Ezoic has helped you solve this problem in reality vs. your current hypothesis of “this is something to fix” based on arbitrary data.
- 84% of visitors are mobile
- 74% of those visitors are on a fast 4G (or better) connection. That equals 63%
- 63% of mobile visitors and the remaining 16% of desktop users all have a fast connection. That means 79% of your website visitors have a fast version of your website.
Using Lighthouse — which is the tool Google says they use as a proxy for actual speed and how they said they would theoretically measure a users interaction with site speed FOR SEO — …
79% of your visitors are getting (and Google is “seeing”) a 100/100 pagespeed score according to their tool, Lighthouse...
What about the other 21%, well luckily for us they aren’t negatively impacted according to the data above (average pageview duration etc.). Even on the slowest possible connection, I could simulate on your site inside Lighthouse, those users still get the content in Google’s 2 sec. threshold.
We are constantly learning what levers impact publisher traffic and revenue. Keep in mind that we are looking at tens of thousands of sites every day and that we have a rare view into the ecosystem that few other technology businesses have.
This has allowed us to optimize a lot of things that are hard to do manually. Site speed is something we are fundamentally trying to help publishers forget about.
Aside from the metrics that don’t matter, competitors on your top keywords actually have a lower Lighthouse score using the exact same methodology I used for your site! Here’s the query I used.
[image removed for privacy]
and here is the competitor…
I clicked on a competitor with a brand name on the same page for the query above.
They had a lower score; partially because of issues they have that Ezoic has solved for you.
Again, not really how Google “thinks” about this sort of thing. I promise (and I think it’s very obvious) that they are not ranking pages because of speed. It would be a terrible searcher experience simply to have results for a query to be based on site speed. In fact, it’s likely not even in the top 20 most important factors.
That is the gist of the of what I think may prove helpful to others following this thread. The ability t breakdown your users using Google Lighthouse.
This way, you can actually see how Google and users are seeing speed on your website.
I will also refer to the best practices I shared above as a mechanism for achieving these scores.
If you're having trouble understanding the information above or still are experiencing issues that seem to be directly related to your website, please contact firstname.lastname@example.org and they can dig into the issue directly.