Not bad findings.
So, SimilarWeb says they use publicly available data to calculate their numbers. Stuff like Google Analytics (RT does have it installed on the website), various APIs and ISP information. A lot of that can be cross referenced across themselves to paint a picture of what sites point to RoosterTeeth, what sites get traffic coming from RoosterTeeth, how many requests for certain APIs are made, etc. I'm still not certain how accurate it is, but suffice to say it does paint a picture of how often the site is visited, and that picture isn't pretty. 2M sounds nice, but for comparison...
View attachment 2868556
(BTW, Direct refers to users accessing the site by typing the full URL into the browser)
As for Ubersuggest, I can't really find how their calculating everything, but I'm going to assume it's the same as SimilarWeb and the number discrepancy between the 2.8M and 500k in visits is because Ubersuggest is calculating the traffic "organically". If I had to guess, their numbers are based off of treating identifiable visitors as a unique visitor rather than treating each visit session as a unique visitor (doesn't care if it's seen you already today, you left the site and came back so you're new traffic). Numbers still aren't great, and once again
View attachment 2868582