Google PageSpeed is a collection of tools by Google Inc for improving the performance of a website. It helps a webmaster or developer to measure the speed of his/her website with various parameters.
This tool was developed by Google engineers, Benita and Youssef Al-Hashemi. Benita and Youssef describe their work on PageSpeed at Google Blog. They used their blog as an opportunity to explain how PageSpeed works. Below are two of the audits they did with the Google PageSpeed tools – User Experience and Load Time.
The first of the Google PageSpeed Inspections were done by Google’s product development team called the Quality Assurance Lab. This group, under the supervision of Sander, Michael, focused on user experience because they wanted to know what users were thinking about the website and whether or not it was easy to use. They did user experience audits by using a simple web-based survey instrument that asked users a series of questions. This revealed a number of problems, the main one being that users did not find the navigation easy to understand or navigate.
The second set of Google Pagespeed Insights were performed by Google engineers Benita and Youssef. They measured the load time of the pages and found that many web pages take too long to load. They also measured the PageSpeed insights of Yahoo! Search and Bing, and found that they also had a number of issues that users were having. These two sites are Google’s competitor sites so this was a good measure of what was going on with their traffic.
There were a couple of other factors that influenced pages speed, one of them being the first CPU idle cycle taken each time a new web page was loaded. Pages used the first CPU idle cycle to search the index, download, and index the content, and retrieve the images and scripts. This takes a very long time, especially when pages have a lot of content. Since Google wants you to be able to use their site as quickly as possible, they need to reduce the amount of time taken in the first CPU idle cycle. By reducing the amount of searching that takes place in the first CPU idle cycle they can improve the speed at which your web page loads.
By monitoring the first CPU load time Google found that the average time taken to load the pages was 4.2ms, which is not where they want visitors to go in their first visit! They found that the average landing page load time was also too long. It took visitors too long to decide if they wanted to proceed with the site. They also found that when they did proceed it was usually after the visitor had spent a considerable amount of time reading the information on the landing page. It also took them too long to leave the landing page once they were finished reading.
The reason for the too-long load times is not clear. It may simply be that people don’t interact with the pages in a way that Google would like. For instance, if there is only one side of the site and the user never goes any further, then naturally they will not spend a significant amount of time reading the content on that side. They will instead move on to the next website that does provide more information or interaction. If they had a user sitting on each individual website that provided the user with meaningful content, then it is likely that the search engine results pages would load faster. If Google could measure interactions rather than pages, it is likely that they would make adjustments to their algorithm to improve how pages rank in the search engines.
There are two main ways that Google works to improve the performance of a website. One of these is through the use of scripts such as redirects and landing page redirects. These work by forcing the browser to follow a different path to a specific page when it attempts to get it to load. There is however a limitation to the effectiveness of these scripts and one of the reasons that they have become such a hot topic recently is because of Google’s change to prevent the use of certain scripts on the landing page.