Your measurements of bandwidth would be accurate if we were still talking about the web in 1999.
However, your math doesn’t take into account that almost every site out there is using both synchronous and asynchronous requests and that the server really does make a lot of difference.
Web sockets allow multiple, asynchronous requests, even on a 3G connection and actually can result in downloading more total data than xKb/s.
Additionally, Apache and Nginx both support gzip delivery by default, IIS supports this as well as some other compression methods. Certain mime types are gzipped by default – like images with PNG, JPEG / JPG and GIF extensions.
Apache can also be configured to handle the synchronous requests while Nginx handles asynchronous resulting in less delay in some types of apps.
The idea that how fast someone’s site loads not taking into consideration the server configuration and this being somehow irrelevant to “stakeholders” is ridiculous and incorrect.
Also, “stakeholders” isn’t clearly defined in your article, but I’m assuming you’re referring to those who own or have interest in the site.