it's actually very easy to measure. the audience here is close to the 1%, so your usecase is already discarded in the grand scheme with a 2% error margin when measuring behavior for the large population.