I had to go back and re-read but couldn't find anything relating to it: why did you assume they used hours of usage as a metric versus the number of visits to the site?