Approximating the Distribution of Broadband Usage from Publicly-Available Data

If a broadband service provider imposes a monthly usage cap of 250 gigabytes (“GB”), how many of its customers would hit the limit? What if it were 200GB? 100GB? 50GB? Oddly, while we can look up on a broadband map what type of broadband is available at every address in the United States, we have no off-the-shelf answers to many basic questions about Internet usage levels.

In this paper, Ford attempts to provide some rough guidance on how Internet usage varies across users, and do so using publicly-available information. His calculations are based on only two data points that, when combined with the pattern Internet usage is known to follow (that is, the statistical distribution of usage), permit the full pattern of usage levels across connections to be approximated. A check on the accuracy of this approximation is conducted using other publicly-available data. Finally, an example of how to use this information, drawing from claims made by Comcast about usage levels and caps, is provided.