HP/Microsoft's study didn't include migration to a newer version of Windows. Considering Microsoft will stop supporting Windows XP as early as next year, not including the upgrade cost for a Windows migration makes no sense, unless they were deliberately trying to be misleading.