PageSpeed Secrets: What is the Difference between Loading Time and Display Time?
Google PageSpeed Insights Secrets
It can be said in advance that there is a very good and also a serious difference. Above all, when it comes to Google PageSpeed, it is never about the loading time, but only about the display time. Are you sure you're irritated now? Don't worry, if you have read and understood this article, there is clarity in any case. However, this article should also give you food for thought, because this article will most likely throw your previous knowledge about optimizing a website overboard.
For a better understanding, we must first discuss the intentions of Google PageSpeed and where the limits of what is possible lie. While the name PageSpeed implies that it is about speed, or PageSpeed, it depends on how you define PageSpeed. If Google uses PageSpeed to score, then this requires constant measurement conditions and that is where the limits lie. There are no constant measurement conditions if you want to assume that PageSpeed would measure the loading time. As is well known, the distance between 2 servers and the performance of servers varies greatly, so that these two significant factors alone allow the safe assumption that PageSpeed can never measure the loading time. Otherwise, every less powerful server and/or remote server would inevitably be disadvantaged in the measurement.
So if PageSpeed cannot measure the loading time at all, what does PageSpeed measure and are there any constant measurement conditions at all?
Yes, there are constant measurement conditions, but such a measurement has nothing to do with the loading time. This type of measurement, which is also operated by PageSpeed, is not about loading, but about the display or the display time. The actual loading process and the time required for loading does not matter for PageSpeed with one small limitation. PageSpeed is therefore exclusively about the display time, which is defined by the fact that it begins after all data has been loaded and ends at the point when the loaded data is completely displayed in the browser. After all, the data first has to be processed after it has been loaded, with the Stylesheet (CSS) playing a major role and, depending on the scope, having a significant influence on the display time. The coveted CSS frameworks such as Bootstrap and other frameworks are therefore anything but advantageous because the amount of CSS would be sufficient to format at least 10 themes.
If it was previously described that the display time could be measured under the same conditions, then this type of measurement also has its limitations. You have to know that the display time is not measured by Google or PageSpeed, but by users who use a Chrome browser. With the User API included in the Chrome Browser, this API not only collects a vast amount of data, but also uses a standardized process to measure the display time in order to transmit it to Google. However, this is not a conspiracy theory, but can be read in the publicly available documentation.
The aforementioned restriction, which is associated with a conditional measurement tolerance, results from the fact that the display time is affected to a not inconsiderable extent by the performance of a device. It's not so much about the differences between desktop and mobile devices, but rather about their different performance, especially in the case of mobile devices. An iPhone/Android smartphone that is 5 years or even older cannot process CSS as quickly as a current smartphone, which inevitably results in more or less large differences in the display time. It should therefore not be surprising if a manual measurement with PageSpeed Insights produces a different measurement result.
If you now internalize the consequences that Google PageSpeed cannot measure the loading time, let alone want to, you must now realize that a high score does not mean that you have a fast website. The PageSpeed score only says that the display behavior has ideally been optimized so well or less well that the loaded data can be displayed quickly or less quickly in the browser. While Google has good intentions for PageSpeed, the purpose of PageSpeed is highly misleading, which is especially true for shared hosting users. PageSpeed's methodology makes them believe that they have a fast website that is, however, extremely slow. So it can't be said clearly enough. No amount of optimization in favor of the display time can improve the loading time. This requires completely different methods for which no expensive XXL hosting is required.
With minimal effort, it is easily possible to significantly outperform a high-end dedicated server with low-budget shared hosting in terms of loading time. What you need for this is available at the same price as conventional hosting with Apache or nginx web servers. All you need is an HTTP cache like LiteSpeed's LScache. An almost 100% Apache compatible web server that is already pre-installed in numerous hosting offers. Thanks to numerous free plugins, you not only have the prerequisite for the best possible loading time, but also optimization functions tailored to the cache that are second to none. More and faster is then no longer and at the same time describes the top of the food chain.
Yes, there are constant measurement conditions, but such a measurement has nothing to do with the loading time. This type of measurement, which is also operated by PageSpeed, is not about loading, but about the display or the display time. The actual loading process and the time required for loading does not matter for PageSpeed with one small limitation. PageSpeed is therefore exclusively about the display time, which is defined by the fact that it begins after all data has been loaded and ends at the point when the loaded data is completely displayed in the browser. After all, the data first has to be processed after it has been loaded, with the Stylesheet (CSS) playing a major role and, depending on the scope, having a significant influence on the display time. The coveted CSS frameworks such as Bootstrap and other frameworks are therefore anything but advantageous because the amount of CSS would be sufficient to format at least 10 themes.
If it was previously described that the display time could be measured under the same conditions, then this type of measurement also has its limitations. You have to know that the display time is not measured by Google or PageSpeed, but by users who use a Chrome browser. With the User API included in the Chrome Browser, this API not only collects a vast amount of data, but also uses a standardized process to measure the display time in order to transmit it to Google. However, this is not a conspiracy theory, but can be read in the publicly available documentation.
The aforementioned restriction, which is associated with a conditional measurement tolerance, results from the fact that the display time is affected to a not inconsiderable extent by the performance of a device. It's not so much about the differences between desktop and mobile devices, but rather about their different performance, especially in the case of mobile devices. An iPhone/Android smartphone that is 5 years or even older cannot process CSS as quickly as a current smartphone, which inevitably results in more or less large differences in the display time. It should therefore not be surprising if a manual measurement with PageSpeed Insights produces a different measurement result.
If you now internalize the consequences that Google PageSpeed cannot measure the loading time, let alone want to, you must now realize that a high score does not mean that you have a fast website. The PageSpeed score only says that the display behavior has ideally been optimized so well or less well that the loaded data can be displayed quickly or less quickly in the browser. While Google has good intentions for PageSpeed, the purpose of PageSpeed is highly misleading, which is especially true for shared hosting users. PageSpeed's methodology makes them believe that they have a fast website that is, however, extremely slow. So it can't be said clearly enough. No amount of optimization in favor of the display time can improve the loading time. This requires completely different methods for which no expensive XXL hosting is required.
With minimal effort, it is easily possible to significantly outperform a high-end dedicated server with low-budget shared hosting in terms of loading time. What you need for this is available at the same price as conventional hosting with Apache or nginx web servers. All you need is an HTTP cache like LiteSpeed's LScache. An almost 100% Apache compatible web server that is already pre-installed in numerous hosting offers. Thanks to numerous free plugins, you not only have the prerequisite for the best possible loading time, but also optimization functions tailored to the cache that are second to none. More and faster is then no longer and at the same time describes the top of the food chain.