And most companies vastly overestimate their data, and believe it to be "big", when it could be trivially handled decades ago by server-grade hardware.