Almost every crawl error on our domain came from a site that didn't exist yet

Search Console said 17% of Google's requests were failing, on a site with nothing indexed. The errors were real — they just belonged to a different site. And the most convincing number in the investigation meant nothing.