if I interpreted the source of the article correctly, then the article makes a big mistake, they count the domains that change from godaddies nameservers. If you don't host with godaddy it wouldn't show in the statistics. Check the link to which the article refers.