So, you’re serious about your page speed and want to run some tests. Excellent! Testing your page speed will let you know how well your website is performing.
Even better, Google PageSpeed and GTmetrix give really useful feedback. You can use this information to fix the issues internally or by hiring a page speed optimization consultant to look after it for you.
By working on your page speed, you’ll positively impact your Google rank and reduce the number of people leaving your site after a few seconds.
If you still haven’t decided what tools to use, or maybe you’re not sure how to conduct a speed test, check out our article titled How to Speed Test Your Website.
Our favorite testing tools are Google’s PageSpeed Insights, Pingdom Tools, GTmetrix, and WebPage Test.
But before you begin, there are seven key things to keep in mind. Not following these tips will give you a headache and cost you time. Many first time testers become puzzled when they notice wildly different results from test to test and tool to tool.
Save yourself the confusion by going through the tips below:
1. Testing Tools Examine Pages Not Websites
This might be obvious to some, but it’s still worth mentioning. All page speed testing tools look at the load time of an individual page rather than an entire website. When you test your homepage URL, the results you get are unique to your homepage and nowhere else.
We like to look at Google Analytics to find the top traffic pages and prioritize those. We also talk to the business to better understand the most important pages in terms of revenue generation.
2. Always Test the True URL
What the hell is a true URL you may be asking.
A true URL is just the URL your server, and
- https://www.yourwebsite.com
- http://www.yourwebsite.com
- https://yourwebsite.com
- http://yourwebsite.com
All of the above formats are correct and your website could be using any one of them. Have a look at the images below to see two different formats in use.
You can see that Load Labz uses the https://www. structure while PageSpeed Insights went with https:// dropping the www.
As I mentioned above, both are correct. Just make sure to visit the page you want to test and test the URL exactly as it’s displayed in the browser.
Errors occur when people type their web address into a page speed tool. If I just typed www.loadlabz.com into Google PageSpeed, my results would be inaccurate because that is not the correct URL for my website. Our correct URL is https://www.loadlabz.com.
The URL still works, but my results will be much slower because my server has to redirect www.loadlabz.com back to my true URL.
The mechanics behind all this is a bit complicated so
3. Set Your Location Before Testing
This is important if you want to use tools other than Google’s PageSpeed Insights. You will have the option to pick a location if you test your website on Pingdom Tools, GTmetrix or WebPage Test. PageSpeed Insights does not have a location option.
Note: You must be signed in to change location on GTmetrix. Luckily, it’s free to set up an account so I recommend you do. You will also get queue priority over those who test without an account.
See the image below to find the option for setting test location on GTmetrix. It only appears when signed in.
When performing a test you want to consider two things. Where are your target visitors based and where are your servers based?
Testing for Visitor Location
You want your visitors to experience a quick loading website so they don’t get bored and just bounce. A recent study reported that 40% of visitors will leave a page if it takes longer than three seconds to load. Ouch!
So, match your test location to that of your target audience. Now you can see the wait time they experience.
If you are trying to reach folks in the US, then set your location to the US. If you’re trying to reach people worldwide, do a number of tests using each available location and see how your times vary.
Improve Server Location if Necessary
The greater the distance between your server and your visitors, the greater the time they will wait for a page to load. Contact your hosting provider or web developer to make sure your website is hosted on a server close to your target audience.
You will notice your results getting worse as you choose test locations further away from your server. For example, Load Labz is hosted in Amsterdam so if we test from Singapore our load times are negatively impacted.
This is not an issue as we are not trying to reach people in Singapore. However we do have clients in the US, so we might consider moving our server over there in the future.
Implement a Content Delivery Network (CDN)
This is slightly off-topic but worth mentioning. A content delivery network can help you improve load times when targeting visitors a significant distance from your server. Implementing a CDN will not solve all your problems but it definitely helps.
We recommend Cloudflare as they have a decent network of servers, decent speed optimization features, and allow you to set up one free account.
In short, be mindful of location when conducting a test and always keep your settings consistent with each batch of testing. And remember, Google PageSpeed doesn’t offer an option to change location while Pingdom Tools and GTmetrix do.
4. Conduct Multiple Tests on Each URL
You’re likely going to get slightly different results every time you test a URL. The biggest difference tends to between the first and second test and this is due to a thing called caching.
Let’s start with a quick explanation of caching. Stick with me on this because caching is crucial to a fast loading site.
What is Caching?
When you click on any link to visit a webpage, you set off a chain of events. Your browser sends a request to the website’s server asking it to fetch that particular page, then the server has to do some calculations to retrieve the page and all its parts. You begin to see a page appear on your screen as this data comes through.
The calculations your server has to do can take time depending on the complexity of a webpage. Things like plugins, widgets, code quality, images and a bunch of other things add to the server’s workload.
Caching tells the server to store the page it just put together rather than do fresh calculations every time a vis
Once the page is cached, the server will just retrieve the page it prepared earlier in a fraction of the time it takes to do all the calculations.
It’s kind of like the way a cookery show will whip out a beautiful dish they made earlier rather than make the viewer wait for one to cook.
And it’s not just your server that can use a cache to speed things up. Your browser also keeps a cache of every website you visited. It stores important resources like stylesheets, fonts, JavaScript files and other things. Together, server-side caching and browser caching
I just want to give an overview of caching rather than get too technical, so I’ll leave it there.
So, How Does Caching Impact my Page Speed Testing?
The first test you run forces the server to do the calculations and build the HTML file. In other words, it will not serve the cached version.
This will show up in your results as
And for this reason, you always want to run a test, wait 30-60 seconds, and run another. Do this about three times and you will get a much clearer picture of your page speed.
5. You Don’t Need to Score a Perfect 100
Yes, I said it. Scoring 100 is not the goal. The goal is to improve the load time of your website. It’s your load time that impacts search rank and it’s load time that makes your visitors yawn and move to your competitor.
These tools are great, but never forget you are striving to improve user experience and your position in Google’s search results. Your score on PageSpeed Insights, Pingdom, and GTmetrix is just a guide and nothing more.
To prove this, I tested the page speed score of each tool using their own tool. I was hoping to enter some kind of page speed Inception adventure. That didn’t happen, but the results I got back were still interesting.
You can see from the results, even the testing tools can’t achieve a perfect score on their own flipping test. Just focus on increasing your load time without breaking the bank or your website and you should be fine.
6. Different Tools, Different Results
Each of the four big testing tools
Pingdom Tools does call out “Load Time”, while GTmetrix calls out “Fully Loaded Time”. They all use different terms because they all have slightly different ways of interpreting what load time actually is.
Pingdom Tools looks at
Considering all the jargon above, it’s easy to understand why people just fixate on the big score at the top of the page. It’s much easier to understand an 80 or an A or B.
Just Keep it Simple
To keep things simple and save your sanity, just pick one tool for measuring load time. At Load Labz, we think the “Fully Loaded Time” in GTmetrix offers the quickest and easiest to understand load time results.
7. Log Everything!
Let’s say you test your WordPress homepage on Google PageSpeed and a whole load of recommendations show up. You might decide to go it alone and fix a few errors using a plugin.
Good call, things like caching and minification can easily be fixed with a decent plugin.
So, you install the plugin and start messing around with the different options. You check some boxes and uncheck others not really knowing what each one does.
Then you test your homepage again to see if it made a difference. Why isn’t it in the 90s or scoring an A, you ask yourself.
You read a few articles and decide to try another even better plugin, maybe you pay for one. Then, you follow an online guide this time and do exactly what the author did on their website.
Finally, you test,
The above scenario is very common. To fully understand what works and what doesn’t, you need to keep a log and test your page multiple times after every. single. change.
It’s time-consuming but very revealing. You will notice which options have the biggest impact and which made things worse. Without a
How to Use a Log
Get a base score by testing the website ahead of any optimization work. You might use metrics like Load Time (Pingdom Tools), Fully Loaded Time (GTmetrix), and Speed Index (Google PageSpeed).
You then want to run a test after every speed optimization change you make. Once you download and activate the plugin, run multiple tests and log the results.
If you check a box in the plugin, run two tests and log the results. If you optimize images, run two tests and log the results. You get the idea.
Keep doing this and you will begin to see what’s having the most impact. You will also be able to identify any changes causing more harm than good. Below is a simple log sheet we use when testing smaller websites. Click it to grab a copy.
One other thing to mention when testing and logging results
You should clear your cache after each change when optimizing your site. This means the first speed test you conduct will not pick up your caching system. As a result, you will see a slower load time.
In Tip #4time-consuming I mentioned the need to do multiple passes on the same URL. You should run at least two tests with about 30-60 second intervals to get a good sense of your true page speed. The first test after clearing your cache will always be slower than the previous.
The process is as simple as this: make change >> clear caches >> run test #1 >> log results >> wait 30-60 seconds >> run test #2 >> log final results. Then move on to the next change.
Thank you
So, that’s it. If you keep these seven points in mind you will conduct much better speed tests.
If you have any questions just use the comment section below or feel free to contact us directly with a specific query.