
In the framework of the project “Let's make the Internet faster”, Google has published some statistics on the size, amount of resources and other indicators of the pages on the World Wide Web. Statistics were collected from a sample of several billion web pages in the process of analyzing and indexing the search engine.
When processing these pages, the algorithm took into account not only the main html pages of the sites, but also attempts were made to detect and process other resources placed on the sites: style sheets, scripts and images.
The main parameters of the pages
- An average webpage size of 320 KB (when transmitted over the network)
- Only two thirds of the compressible data were compressed enough.
- 80% of all pages load 10 or more of their resources from one server
- The most popular sites can eliminate more than 8 http-requests per page if they combine all the scripts on the same host into one and all style sheets on the same host into one.
Main disadvantages
- All resources were processed by Googlebot, so they could fall under the restrictions of robots.txt. Some sites (for example BBC) block CSS and JS.
- Some sites may provide a different set of resources for Googlebot and for regular users. For example, until recently, Google’s own servers kept CSS and JS uncompressed for Googlebot, but compressed them for regular browsers.
- If page resources on the server are different for Internet Explorer or Firefox, they will not be visible in WebKit.
- The selection of pages for processing is not uniformly random and impartial. For example, pages with a higher PageRank value are more often included in this sample.
During the analysis, popular sites were considered separately compared to all other sites from the sample. As it turned out on popular sites, the average number of resources and GET requests per page is less than the rest, while they use more unique host names, but fewer resources per host.
The average page of the site from the top was less than 8 kb in transmission over the network, but more than 100 kb in uncompressed form, while they were compressed much worse, all this is due to the fact that the resources on such sites were initially compressed with higher quality.
Pages of top sites contain an average of 2 unique images smaller than regular ones, the size of which, like the size of external scripts, is slightly smaller, while the style sheets on the top sites are one and a half times larger than the styles of regular websites.
Analyze the statistics yourself, as well as get acquainted with the project Let's make web faster, you can
at . You can also find recommendations on the site that will make your applications faster, stay up to date with the latest news on web productivity, get acquainted with various tools that can help improve the performance of your site.