"Hi Joshua, Thanks for stopping by and good question. I think you're looking at the data from a user-centric perspective, audience overlap. This data is publisher-centric. I.e. amongst an overlapping publisher (websites) base what is the penetration rate of each bug. Implicit in that is the assumption that there will almost always be overlap. So it breaks down to be the sum of # of unique sites that these 10 bugs were found on and the relative distribution of each. Hope that helps. ;dc" - David Cancel