I think you may be calibrating for the wrong thing in terms of sampling. The idea that this is "just one building" doesn't mean that it isn't a representative sample: this one building represents hundreds of leases, if not a few thousand. If the aggregate value of those leases is down 75%, that means something.