On 17th March 2017, Google announced its new open source JPEG encoder “Guetzli“, meaning cookie in Swiss German. The new encoder allows for up to 35% compression without significant loss in quality.
The new encoder is aimed at making browsing faster as more compressed images mean smaller file sizes and thus lower load times while still being compatible with the existing JPEG standards, internet browsers, and image processing applications. The encoder works at the quantization stage, where most of the loss in quality occurs.
It balances file size and loss of data by using a search algorithm to adjust the differences in psycho-visual modeling between itself and that of the standard JPEG format. In experiments with human viewers, it was found that Guetzli images were preferred over the ones produced by lib jpeg at almost the same file size. So it seems that the new encoder can produce good quality images at the lesser overall file size.
This is aimed at reducing the overall load time and bandwidth cost, particularly for mobile users as data usage and time is taken to load a page is very important for mobile web surfing. It is also hoped that it will soon be widespread and that graphics designer and webmasters would shift to this shift to using this format for websites loaded with images. This also helps Google indirectly.
As they cannot control where all the images are generated at, they can control how they are stored and read by computers. As most of this data ends up on their clouds, it would cost them less to store them.
Also, Google makes most of it earnings through online advertisement. By reducing load times and data usage, the user will be able to open more pages, each with an ad from Google thus helping them generate greater revenue in one of their primary business.