Is this including non-SOA domains? If so there's a huge potential for over counting even if wildcards are ignored. For example, many isps and server hosts will

Is this including non-SOA domains? If so there's a huge potential for over counting even if wildcards are ignored. For example, many isps and server hosts will serve systematically generated domain records for individual server instances and even home networks. (often these take the form of ip.ispname.net or similar) even more troubling is that these could potentially resolve correctly even if there is no actual website there. Without a more rigorous method like looking at SOA I don't see this being useful as an absolute metric.