In this paper, we investigate two types of IQ-data transfer methods for cloud MIMO network operation. They are termed 'after-precoding' and 'before-precoding.' We formulate a cloud massive MIMO operation problem that includes the best IQ-transfer method and beamforming strategy (beamforming technique, the number of concurrently receiving users, the number of used antennas for transmission) to maximize the ergodic sum-rates under the limited capacity of the digital unit (DU)-radio unit (RU) link. Based on our proposed solution, the optimal numbers of users and antennas are simultaneously chosen. Numerical results confirm the sum-rate gain is greater when adaptive 'after/before-precoding' method is available than when only 'after-precoding' IQ-data transfer is available.