About GTmetrix, PageSpeed Insights and other website speed testers
Make sure to read the related articles (bottom of this page) in order to learn more about specific issues you may encounter.
There are endless situations one can encounter while checking the speed of a website on a website speed tester, especially considering that there are many of those. Let's start by stating that not all of them measure the same things, so it is normal to get different results depending on where you test your site. So, first of all, please have a read at this article by Kinsta to know how to properly run a website speed test: How to Properly Run a Website Speed Test
Our key takeaways from what we have learned in years of experience are:
- Each speed tester measures your sites in a different way and using different algorithms, so like we said, it is normal to get different results. For example, GTmetrix is more technical, while PageSpeed Insights focuses more on user experience.
- Each result is always different than the previous one because such tests depend on many things: the current traffic from the test server to your site, the time of the day, the load of the test server, your caching solution, etc.
- Real browsers provide a better indication of your website’s performance.
- The scores are a way to tell you how good a website is performing in general. If there wasn't a score, it would be too difficult for the average user to understand the results. Such scores are a combination of many factors, most of them unrelated to ShortPixel. Since these factors change with each test, it is possible that they affect the overall score.
- The loading time, in general, is better with our plugins. This is what ShortPixel is aiming for, a better image optimization and thus less page size and less loading time. Like we said, looking at the general score can be misleading.
- Website speed testers are run by algorithms, and they are useful, but not perfect. As with every software, they also have bugs, and sometimes they give results which are not correct. Don't take their reports as the absolute truth.
Before checking your website speed
It is important to know that before doing a website speed test, you should flush your cache. Please do it in this order:
- SPAI cache: only if you are using ShortPixel Adaptive Images, go to Settings > ShortPixel AI, and press the button "Clear CSS cache".
- Cache plugins (WP Rocket, LiteSpeed Cache, SG Optimizer, Autoptimize...)
- Server cache (SiteGround, BlueHost, Kinsta...)
- CDN cache (Cloudflare, StackPath...). Please do it from your CDN's dashboard, not through your cache plugin or hosting dashboard.
After flushing your cache, you should check your site with a page speed tester multiple times in order to let all the caches work properly and rule out any false positives.
A bad score does not mean a slow site or a bad ranking
As mentioned above, the scores are a combination of many factors, and many of them have no effect in your website speed. This is especially true in Google PageSpeed Insights (and GTmetrix, as their results are based on Google's); you may think this tool matters the most because it’s made by Google, but the reality is that there’s very little correlation between getting a good score on this tool and having a good rank on Google. The information it gives is either unhelpful for the average user or highly off the mark. In other words, scoring bad in PageSpeed Insights means very little. To illustrate this, check these two famous sites:
- Mercedes Benz: https://developers.
google.com/speed/pagespeed/ insights/?url=https%3A%2F% 2Fwww.mercedes-benz.com%2Fen% 2F&tab=mobile
- CNN: https://developers.
google.com/speed/pagespeed/ insights/?url=https%3A%2F% 2Fedition.cnn.com%2F&tab= mobile
Tanner, from UpCity, gives a very good advice:
When it comes to site speed as a ranking factor, Google isn’t looking for a website with lightning-fast speed, it’s looking for one that can meet a user’s intent. As such, user experience and content should be the focus of any SEO campaign. Site speed is merely a small piece of the puzzle.
Sometimes one will be concerned regarding the TTFB (Time To First Byte); some hosting providers may claim that ShortPixel is taking a few ms of loading time. Good news: you shouldn't be concerned about the time to first byte (unless it's extremely bad). Any image optimizer (and actually any plugin) by definition will take a few ms of time when loading. We can assure you that ShortPixel's plugins are among the most optimized ones, and you can try them out in a sandbox like this one: Sandbox
In addition, we recommend to read this article that Cloudflare wrote about this clarifying this situation: Stop worrying about Time To First Byte (TTFB)