You collect "aggregate" statistics from the open source server by default?

Do you have a whitepaper describing your aggregation methods? I've seen a talk by some Google engineers who worked on Maps and they went to extensive efforts to anonymize drive segments. This is very difficult (there's a surprising amount of subtlety) and while I'm not doubting your company's abilities, publishing such a whitepaper would help dispel any misgivings people might get over knowing that you collect this by default.