Before going into the idea of testing site speed, it’s best to understand, on a more than basic level, what makes a site fast. At its core, site speed is a pretty simple concept. Your site either loads quickly, or it doesn’t. The hard part is in understanding exactly what is causing the site to be relatively slow, or relatively fast. So, the question remains: why does a page load in a certain amount of time?
What happens when I load a page?
When you type a URL into your browser, your browser looks for the site connected to that URL and then brings it to your screen. On a more detailed level, your browser shoots out a request for the code connected to that URL, and once the network pulls the resources from servers and fulfills the request, it sends the code back to your browser. At this point, your browser processes the code and puts it up on your screen. While the browser is processing the code, it executes queries (like requests) in a certain order. These queries tell the browser what to load in order and how to load everything. For example, the structure of the page (locations of headlines, photos, sections of content) will probably load first, then the functionality, then the styles and content. When you begin to look at exactly what is being loaded in any particular order, that is where you begin to see the pieces that make up the overall load time of the page.
All pages are not equal
There are a lot of things that can affect page load times, but there are some that affect it more than others. Assets on the site can make a gigantic difference. By assets, this refers to anything like fonts, photos, or videos. Photos are the big ones, especially in the tour industry. If your home page is just full of photos over 1 MB in size, your overall loading time is going to be affected, because you have multiple large assets on the page.
Loading time can also vary depending on functionality. If the home page of your site is a simple page with content and photos, and no fancy features, chances are it will load a bit faster since the nice UX (user experience) functionality of some features won’t have to load. In the terms of UX, finding a good balance between functionality and speed is very important because you want the site to be interesting so that users interact with it, but you don’t want it to take so long that they bounce from the site.
External resources can also affect loading time. As the page is loading, once it gets a query to load something from an external resource, it needs to reach out and pull from a third party, which can add time. An example of this is Google Analytics or Tag Manager, which has you add a tracking tag to your site in order to collect data.. While pulling most of the code for the site from the network, when the browser comes across the tracking tag, it then also needs to reach out to Google. Even though it happens fast, it adds to the loading time overall and let’s say you have 4-5 different tracking pixels or codes on the site, that can add up.
Another differentiating factor is perspective. Accessing a site from a different browser, a different device, from a different country, and from different networks (WiFi or Service Providers) can all affect the loading time of a page. For example, using an old iPhone 4 versus the newest iPhone will likely load the page in different times. In terms of service providers, let’s say you live in a city with 2 service providers but outside the city only 1 of those service providers gets a good signal. If you don’t have that service provider and then access the site from outside the city, you will likely get a different loading time than if you were in the city.
How to test page loading times
If you have a website you’ve probably heard of Google Pagespeed Insights(opens in a new tab), as it is the most common, beginner-friendly site testing tool. While it is a good tool, the most common mistake is getting caught up in the overall number and just assuming the site loads poorly or well depending solely on that number. When using a tool like Pagespeed Insights or GT Metrix(opens in a new tab), the test pulls those scores from multiple elements of the page—not all of them relating to speed. For example, it is very possible that a page could load in 3 seconds and still get a bad overall score on those tests based on certain items. If your site loads fast, but is not secure via an SSL, you would see a lower score than if your site was fast and secure. Instead of looking at overall scores, when using those 2 tools, you should focus on these specific numbers:
- Time to First Byte (TTFB): This metric largely reflects the quality of your hosting servers. It measures how quickly it takes for the server to begin sending data to a user trying to load the page.
- First Input Delay (FID): This number tells you how long it takes from when a user clicks on a link or button to begin to load the correct reaction on the site. For example, when clicking on a button to bring up a booking window, you want it to begin loading as soon as possible rather than having a user on the site waiting at a blank screen for a while.
- Largest Contentful Paint (LCP): This stat is a great indicator of loading time in terms of what the user sees. The LCP time is how long it takes to load the largest asset on the page – most commonly something like a hero image or video header.
- Cumulative Layout Shift (CLS): This one has nothing to do with speed! It is a score based on how many elements on the page shift around while it’s loading. Have you ever been loading a site, saw the button you want to click on, and then right as you click the page moves and you end up clicking on something else? That’s a layout shift, and obviously, search engines want to see good scores here because they don’t want users to land on wonky sites – and you don’t want to have one!
Another tool we recommend is WebPageTest(opens in a new tab). This is by far the most detailed of the site speed testing tools that we commonly use at Tourism Tiger. There are a few reasons why we prefer the results of this tool over the other two:
- First and foremost, WebPageTest runs page tests using real devices where you can even choose the location. Pagespeed Insights and GT Metrix use virtual machines when they run their tests – meaning the connection is artificial and therefore not entirely reliable. With WebPageTest, you can test on real device connections to get a strictly realistic idea of how your site loads.
- WebPageTest also runs a number that the other tools do not have: Speed Index. This number is a realistic loading time of your site above the fold. Considering that when users load sites, they only see above the fold assets loading, the Speed Index gives you a great idea of how users are seeing your site as it loads.
- The recommendations based on results are much more detailed in this test. While they may be a bit technical, most experienced developers will be able to look at these reports and get a good idea of specific fixes to be made. On a more non-tech-savvy level, something you can look at and understand easily are the waterfall charts(opens in a new tab) that they give with the tests (GT Metrix shows one as well). These charts give a good visual representation of what loads at what time, and how long each item takes. For example, let’s say your LCP is really long, like 4 or more seconds. Chances are, on the waterfall chart, you are going to see one horizontal bar much longer than the others, and it will likely be correlated with whatever image or video you have as the LCP on the page.
Don’t stress over fully loaded time either
Now that you know some key elements to look at when testing your site speed, you may still be tempted to think of the fully-loaded time as a key statistic. While it is important to look at, keep in mind that this stat takes into account the entire page, meaning that even assets loading below the fold(opens in a new tab) (which users won’t see) go into this number. It is much better to have a 2.5-second LCP or 3 second Speed Index with a 10 second fully loaded time, for example, rather than having a 4-second LCP and Speed Index with a 5 second fully loaded time. Also, if you’re worrying about search engines caring about fully loaded time, don’t. Google will place more ranking weight on the FID, LCP, and CLS by adding them to their list of search signals(opens in a new tab). When you run speed tests, don’t focus on overall scores, but make sure to look into those 3 metrics. With this move by Google, it’s safe to say that they are continuing to target pages with realistic page speeds (based on what users see as it loads) for better ranking.
Do your site pages already look good for those search signals? Make sure they’re mobile-friendly and secure as well. That covers most of the technical signals for Google and other search engines. Once you have your site set on the technical side, make sure you are looking good with SEO. We offer Keyword Tracking(opens in a new tab) and Backlink Profiling(opens in a new tab) services to help with on and off-page SEO.
