Google PageSpeed ​​Insights: Google’s tool for analyzing the loading speed of web pages

You know the adage: better a fast website than a slow one. Except that the loading time of web pages is not only a matter of preference: it is also a question of performance. Because the longer a page takes to display, the more the user’s frustration increases, and the more the bounce rate climbs — up to +123% for ten seconds of loading — which has a direct effect, although marginal, on its positioning in the SERP.

Suffice it to say that this metric plays an essential role in the success of a website. Fortunately, there is a tool to measure this speed and get recommendations for improving it: Google PageSpeed ​​Insights. How does it work, exactly?

What is Google PageSpeed ​​Insights?

PageSpeed ​​Insights is a tool launched by Google in 2018, completely free, for anyone who develops, administers, or optimizes websites. By indicating the desired URL in the dedicated field, it is possible to launch an analysis of the page to learn from it about its loading speed, in its desktop or mobile version. The goal is to be able to correct the blocking points in order to improve the display time of the page concerned.

Google PageSpeed ​​Insights is not the only tool on the web to offer this type of audit: many other solutions work on the same principle. However, the interest of Google PSI (for those close to you) lies in the wealth of information provided: speed score, laboratory data (from Lighthouse) and field data (collected via the Chrome UX Report from real user experiences ), diagnosis, and opportunities for improvement. The tool thus makes it possible to generate a very comprehensive report.

In addition, Google PageSpeed ​​Insights prioritizes the mobile version of a page, when it exists. Knowing that the search engine’s Mobile-First Index values ​​the mobile optimization of websites, this analytical focus takes on its full meaning.

Loading Web Pages: Does Speed ​​Really Matter?

We can legitimately wonder about the interest of a tool like Google PageSpeed ​​Insights if we ignore the importance of this metric.

To understand this, ask yourself a simple question: as a user, would you stay on a web page that takes more than five seconds to display? The answer is probably “no”. In the world of the Internet, speed is a virtue, and webmasters are encouraged by search engines to optimize their sites so that they load as quickly as possible, ideally in less than three seconds (according to John Mueller, Senior Webmaster Trends Analyst at Google).

A study conducted by Google in 2017 showed that the increase in the loading time of a page is accompanied by a proportional increase in the bounce rate:

For its part, in 2019, backlinko analyzed five million pages (desktop and mobile versions) to arrive at the following statistics: on average, it takes 10.3 seconds to display a desktop page in its entirety, and 27.3 seconds for a mobile page.

However, everything leads us to believe that Internet users’ requirements in terms of speed have increased since these studies, in particular with the development of mobile Internet. It turns out that mobile users are even less patient than Internet users: 85% of them expect pages to display faster on a mobile device than on a desktop computer. A UK study even showed that 14% of shoppers want pages to display instantly on mobile! (You can find these numbers in this article.)

That’s not all: beyond its impact on the user experience, the loading speed of a page also influences its positioning in the SERP. This metric is even a key factor in SEO. This is because Google’s algorithm works on a simple equation: providing the most relevant answers while requiring minimal effort from Internet users. In this sense, the loading time of a page can affect its ranking in several ways:

• if users leave the page within seconds of arriving (which affects the bounce rate),
• if the constituent elements of the page take too long to load,
• and whether to wait too long to interact with the page (and receive a response).

Since the implementation of Core Web Vitals, Google has gone even further in its measurement of web page performance. These three factors take into account technical issues likely to have a strong impact on the user experience: the time it takes for the heaviest element of the page to be displayed, the delay between the interaction and the response of the browser, and the positioning of clickable elements on the page.

This is where Google PageSpeed ​​Insights comes in, a tool for measuring the speed of a web page and identifying the brakes that are slowing it down.

What data does Google PageSpeed ​​Insights provide?

After performing the audit of a web page, Google PageSpeed ​​Insights provides a number of data. Let’s see what information you can find in the final report (which, let’s remember, concerns a single page ).

The performance score

The score takes the form of a score assigned by Google PageSpeed ​​Insights to a web page according to its technical performance, between 0 and 100. This score is determined from statistics from the tool lighthouse. Reading the score is as simple as possible, being associated with a color code:

• In green, the scores range from 90 to 100: satisfactory speed.
• In orange, scores ranging from 50 to 89: acceptable average speed.
• In red, the scores range from 0 to 49: slow web page.

The statistics used to calculate this score are six in number:

• First Contentful Paint
• Time to Interactive
• Speed ​​Index
• Total Blocking Time
• Largest Contentful Paint
• Cumulative Layout Shift

Each of these indicators has its own score which measures a very specific criterion. The average of the six scores is used to determine the final mark, after applying a particular coefficient (for example, the Total Blocking Time counts for 30%, while the First Contentful Paint counts for 10%). The detail of the calculation taken from Lighthouse can be viewed by clicking “Show Calculator” below the speed score. You can play with the values ​​to see how these changes affect the final grade calculation.

For obvious reasons (the importance of the concept of “rating”, instinctive apprehension of the color code, etc.) this performance score takes on disproportionate importance in Google PageSpeed ​​Insights — sometimes to the point of obsession. It is, however, an indicator to qualify: rather than seeing it as a definitive judgment, it must be used like a compass to go in the right direction.

As such, it should be noted that the statistics fed back from the Lighthouse API are established from a controlled environment, which reproduces the display conditions of a page, taking into account parameters defined in advance. These statistics are also called “laboratory data” because these conditions simulate the display of the page on a desktop or mobile, without taking into account the real experience of the user. This means that this method, as effective as it is, does not identify the slowdowns of a page in real conditions. To make an analogy, it’s the same difference that exists between testing an airplane in a simulator and flying it for real!

Essential Web Signals

This assessment incorporates the famous Core Web Vitals. These signals allow Google PageSpeed ​​Insights to analyze in detail the user experience relating to a specific web page based (and this is essential) on real data from Internet users. What are these signals?

• The display of the first element of the page (First Contentful Paint)
• The display of the largest element on the page (Largest Contentful Paint)
• The time elapsed between the interaction and the response of the browser (First Input Delay)
• The visual stability of page elements as they display (Cumulative Layout Shift)

The values ​​used here are taken from the Chrome UX Report, which provides metrics about actual website usage by users over the past 28 days. When this data is not available, the Google PageSpeed ​​Insights report makes it clear. Because they are as close as possible to reality, this information helps to understand how Internet users interact with a page, and ultimately to measure its level of technical performance. It is all the more interesting that Google being the first search engine in the world, can rely on a vast volume of data.

The diagnosis

At the end of the report, Google PageSpeed ​​Insights provides users with “opportunities” in the form of a performance diagnostic. These opportunities are all recommendations to apply to optimize the page loading speed, improve the results of the metrics, and thus boost the score. A code combining colors and geometric shapes makes it possible to identify at a glance the improvements likely to generate better results: a red pyramid designates a priority action, while orange square points to a secondary lever. (Note that it is possible to display the audit result by metric type.)

The avenues to explore to improve the loading time of a page can be many and varied. Here are some examples:

• Eliminate unnecessary resources that block the display of the page. These are often scripts (Java or CSS).
• Reduce requests to speed up server response time.
• Optimize images and videos (larger files) by compressing or resizing them.
• Limit the use of 301 redirects.
• Enable Gzip compression for display on mobiles.
• Etc.

Following the diagnostic, Google PageSpeed ​​Insights also details “passed audits”. These are the performance conditions that the page already meets with respect to the expectations of Internet users and the requirements of indexing robots.

Do Google PageSpeed ​​Insights results have an impact on SEO?

This is probably THE question you are asking yourself now that you have fully understood how Google PageSpeed ​​Insights works: to what extent does this data influence the SEO positioning of your page?

The answer must be done in two steps:

1. Page speed is a ranking factor on Google, which means that the engine relies in part on the results of the Google PageSpeed ​​Insights audit to rank pages in the SERP.
2. But, at the same time, the performance score as such is not a natural referencing lever. It is the loading time of a page that counts because it influences the user experience. Therefore, a “green” rating in Google PageSpeed ​​Insights does not in any way guarantee a higher position in search results.

The score, therefore, has no direct influence on the SEO of a website. Its role is elsewhere: to testify to the problems encountered by the user when browsing a page and to allow the webmaster to correct them. This brings us directly to the limits of this tool…

What are the limitations of Google PageSpeed ​​Insights?

Google PageSpeed ​​Insights is a practical and effective tool used by many professionals. However, it has certain limitations that it is important to know, especially about the famous performance score. Many webmasters make this score a word of the Gospel and tear their hair out to reach the highest when the real question to ask is rather: what relevant information does this score bring you?

In reality, there are multiple levers to optimize the score displayed by Google PageSpeed ​​Insights without improving the actual performance of a page, that is to say without modifying the quality of the user experience. This is the case, for example, when touching the source code. However, the goal of the game is not to convince Google of the degree of performance of a page, but to offer a good experience to Internet users by boosting the real speed of display.

As such, it is better to forget the score as such to focus on three points, much more important:

• The recommendations displayed in the diagnosis, are sources of real improvements in the loading time of the page.
• The performance perceived by your users, who don’t care what score you have been given by Google PageSpeed ​​Insights — and who rely on their real experience on your site.
• The loading speed of pages in the mobile version, because this is where most of the technical problem comes into play (according to Backlinko, the same web page takes on average 88% longer to load on mobile than on desktop ). The problem is that a page can rank highly in its desktop version, but show poor performance in its mobile version, which therefore requires specific optimization.

In short, the objective of Google PageSpeed ​​Insights is not to launch a race for the best rating, but to encourage webmasters to improve the loading speed of their pages in order to optimize the user experience. These actions have the advantage of acting positively on Internet users AND on indexing robots. Thus, a page that loads faster tends to minimize the bounce rate, boost the conversion rate, increase the crawl speed of robots and the indexing rate, and ultimately improve natural referencing.