LPWAN enabled networks have a dizzying growth and continue to meet an essential need in the Internet of Things market due to their ability to provide low-cost wireless access to geographically spread-out devices. Consequently, an efficient allocation of wireless resources so as to support a large number of devices becomes a major concern. In this paper we propose an SF assignment approach in LoRa networks paying attention on the traffic load both per Spreading Factor and over the channels. Indeed, our strategy consists in finding a better distribution of the nodes on the SF by orchestrating an effective load balancing. Moreover, the performance of our solution is evaluated under diverse network configurations taking into account the capture effect and the non-orthogonality of SFs. Furthermore, we extended the solution by a transmission power control strategy for overall system energy-efficiency. In addition, we validated some assumptions by full-scale experiments like for the 3GPP path loss model which is used for the first time in LoRa simulations. Our results suggest that Load Shifting leads to better performance in terms of DER (Date Extraction Rate) while guaranteeing good scalability on the network size and density.