Reviewing Popular SEO Page Speed Tools - Why I Recommend WebPageTest
Last Updated -
Fast web page render times are key for SEO success. Google has stated over and over that speed is important.
But you can't know if you are fast or what to address to be faster unless you use tools to measure your page speed. So what tool(s) should you use? And of course my choice is free!
Fast render times are even more important now that most online activity is on a mobile device. This means slower, less reliable network connectivity and slower CPUs.
Surveys continue to show more consumers expect pages to load faster on mobile devices than desktops.
Why is speed so important?
No one likes to wait.
Studies show anxiety levels rise when a person must wait on a web page to render. If it takes too long they will find another answer or solution to their question.
Google calls this bounce rate and a bad bounce rate negatively affects your ability to rank.
What does it mean to load fast and how can you measure your page speed to make sure your site performs well.
Since these tools score our pages I decided to score them back!
Let's see how they do and what tool I recommend you use over all other tools...
Pingdom
Love2Dev Tool Score - D-
This is a popular tool with online marketers, but honestly it is bad, really bad.
The report looks nice and all, but they don't measure much that offers real value. That is unless your server is really, really bad.
You can enter a URL in their free online tool. They let you select from a few locations around the world to see how long it takes your content to reach the client.
And that is the clue, the time it takes your bytes to reach the client computer.
Pingdom only measures time to first byte, which in the scope of page speed accounts for 5% or less of your performance metrics.
It gives this site a score of 86% because of the Google Analytics and other tracking pixel requests. Those all come from external domains and do not have cache headers applied.
Otherwise I would score 100.
They give you a nicely formatted network waterfall. And as you can see from my profile, even including the favicon request my page's assets reach the browser in about 500ms.
My advice is to skip this tool, there are better, free options.
GTMetrix
Love2Dev Tool Score - D
GTMetrix is another popular tool for search engine optimizers. This is another free tool, and like Pingdom it leaves a lot on the table.
The tool does offer a little more insight, but not much.
GTMetrix does run a basic audit of best practices on the page, which is nice. These are all good performance best practices, most I classify as low hanging fruit.
If you score bad in the PageSpeed items then you most likely have a configuration issue with your server or deployment process. It might also be your site's theme, so look in all these areas to improve.
The next tab is a bad sign in my opinion, YSlow.
Don't get me wrong YSlow was great, in its time. But the project has been deprecated for a few years as technology has progressed.
YSlow was created by a team at Yahoo, yeah that company we have all forgotten, as a FireFox plugin you could use to evaluate your page's performance.
They were based on research lead by Steve Souder's team back in the 2007 time frame.
Since then HTTP/2 and other changes to the web have happened, which make some of the YSlow recommendations obsolete.
For example, domain sharding, you should not shard anymore, but use HTTP/2. You should also use smaller files so you can control caching at a granular level.
Like Pingdom the tool really just measures how long it takes your page's resources to reach the browser.
Google Page Speed Insights
Love2Dev Tool Score - C+
This is the last of the popular SEO go to tools. It makes sense this one is offered by the Google search team, so it has to be good, right?
Compared to Pingdom and GTMetrix the Google Page Speed tool is much better. It does not bother with basic time to first byte, but focuses more effort on the rendering cycle, which is the key to good page speed.
The Page Speed tool recently went through an overhaul where it now runs its test using Lighthouse, which I will cover later.
The real value is the Lab Data section. Here you will see:
- First Contentful Paint
- Speed Index
- Time to Interactive
- First Meaningful Paint
- First CPU Idle
- Estimated Input Latency
The real measurement that matters is Time to Interactive, the time when the page is fully interactive. This means when the page has completely loaded all assets, rendered and finished executing JavaScript.
And that my friends is how you measure real page speed, not how lond it takes the server to respond.
All the time up to the point the browser's UI thread is busy loading and rendering page assets.
Google Web.Dev
Love2Dev Tool Score - B-
While the Google Page Speed Insights tool recently migrated to using Lighthouse, they went one step farther by launching Web.Dev.
Like the page speed test, Web.Dev runs Lighthouse over your page. You get a very similar set of data, just more of it.
As a bonus you can view the actual Lighthouse report.
You do not have control over where the test is run. It is run from a Google server, somewhere in the cloud.
Speaking of Lighthouse...
Lighthouse
Love2Dev Tool Score - B+
A while back the Chrome team started working on a site auditing tool, with a focus on page speed.
After a few iterations, including a deprecated Chrome extension, Lighthouse was born.
There are two ways you can run Lighthouse, from your browser and using the Node CLI.
To run Lighthouse from your Chrome browser open the developer tools (F12). Go to the 'Audits' tab.
You will be greeted with the ability to select which audits you want to run and the conditions the tests will execute.
I select all audits, because why not get everything!
Its free after all.
For conditions I like to select 'simulated fast 3G' and mobile. This should execute your page in a more harsh environment. This will reveal where you have issues more than if you run using your desktop and high-speed connection.
For the record, never test using high end equipment, like your developer's machine.
The Lighthouse report is rather detailed. It also gives you a nice filmstrip view of your page loading. This allows you to see what the page looks like to the consumer over time.
Running from the developer tools is very convenient, but I also mentioned the node CLI (command line interface).
This is great because your devops team and developers can integrate it in the site's build and deploy processes.
The tool exercises the page and creates a report in JSON format. You can then use the report data to verify the page is within your allowed performance budget.
SEOs will probably chuckle at the SEO audits because they are so high level. But look at the report this way, it is just another eyeball on those basic tasks, which your developer most likely overlooked.
Also, all these performance tools offer valuable SEO analysis because they are measuring and checking for many technical SEO requirements that most developers and non-technical folks overlook or do not know about.
WebPageTest
Love2Dev Tool Score - A+
Finally, the best page speed testing tool, WebPageTest.org!
If you really want to know how well your site and pages load you will use this test because it offers the most thorough coverage.
And it is FREE!
You can visit the site's home page, enter your url, select where the test should run and what browser or device should be used.
I like using the real phones in the Dulles data center to run my tests because they use a real cellular connection and real phones.
Public instances of WebPageTest are available from data centers all over the world. Because this site is popular in India, I often run tests from the Mumbai instance as well.
Honestly, I don't have the space here to dive into all the data collected by the tool. I focus mostly on the initial scorecard in the top right corner of the report, the times in the table just below the scorecard and the waterfall.
I have been using WebPageTest so much over the years I can glance at these three visuals and just know the basic health of the web page.
Like the Google Page Speed and Lighthouse you get the time to first interactive score as well as the full time it takes for everything, including the content being loaded by the service worker, to fully load.
For the record you want a 3 second or less time to first interactive, even on mobile. Many will say 5 seconds or less, but I have higher expectations.
Why?
Because if you achieve 3 seconds in a known environment then you should be faster than 5 seconds even in the worst environment.
You can also capture the Lighthouse report within the tool, which in my opinion make running Google Page Speed and Web.Dev a bit weak😊💪.
By default, the tool loads a page twice, this gives it the ability to check for properly cached URLs. You can adjust this and run it more times if you like.
You also get a filmstrip view. But even better you can also watch a video of your page rendering!
If you want to get more fancy or stand up you own WebPageTest instance you can. There is an API you can exercise to automate your tests.
WebPageTest is an open source project, which means you can clone it and stand it up in your data center. You can also clone an Amazon EC2 instance or use a container.
Don't worry if you don't understand what I am talking about, your developers and devops folks will.
If you have not guessed by now, WebPageTest is the tool I recommend always using to verify your site's page speed.
It has everything you want. It is regularly updated. It is easy to use and produces reams of data.
You can automate and stand up your own private instances if you want.
Finally, it is FREE!!!!
Monitor Page Speed Over Time
Performance is never a one and done effort. It requires investment over time. Tiny tweaks can have a big impact in either direction.
Sometimes different factors can affect a single measurement. Maybe a router was overloaded, which temporarily increased the network latency. Maybe a high CPU process was running on the client machine creating a longer time to first interaction. Maybe a third party dependency's server was down or unavailable, which created a delay.
This is why I typically do not rely on a single test, but run multiple passes and over time. I look for a consistent pattern to determine the real page speed.
But this can be tedious if you do it by hand. You cannot always remember to run a test every day or even for a large set of URLs. This is why an automated tracking tool or service can be a game changer.
This is what SpeedMonitor.io does for you. Just enter a URL, then some contact information and your first report is off and running!
You can add additional URLs and even track competing URLs. Over time you can see trends and patterns emerging.
If you are actively making tweaks to pages or the entire site you can see how your changes affect the
There are four key metrics, which just happen to map to 4 of the most important metrics in my opinion:
- Time to First Contentful Paint
- Time to First Interaction
- Speed Index
- Time to First Byte
The data is collected from Google Page Speed Insights. Again the best thing is your tests will be run automatically each night for you. Plus since it is the Google Page Speed Insight tool you are getting data Google thinks is important.
Here's the best thing...
It is free!!!
So other than time needed to configure your URLs to monitor there is no financial commitment.
The service is new, so expect improvements and new features over time.
Wrapping it Up
Ok, these are the four most popular page speed testing tools. As you can tell I don't really care for Pingdom or GTMetrix and love Lighthouse and WebPageTest.
If you are an SEO, web developer or business owner you should make Lighthouse and WebPageTest part of your workflow.
If you have pages that are not scoring well, these tools will quickly show you where you are failing so you can take action to improve not only your search ranking, but your ability to engage customers and even make your staff more productive.