I think the formula should be - severity * (how widespread is this) - ease of workaround = priority. So if any of those measures change (e.g. an easy workaround is discovered, it's determined that the bio page that is crashing is also almost never viewed) then the priority should be adjusted. Having just severity without a measure of 'how many people does this impact?' and 'just how badly does this impact them?' seems like it's missing part of the picture.