My guess is that that will work to his advantage -- since it's so hard to tell who is a bot from the raw data, he'll be able to massage his calculations and come up with pretty much any figure he wants, allowing him to claim that Twitter has significantly more bots than claimed