The Internet of Things (IoT) will shortly be undergoing a major transformation from a sensor-driven paradigm to one that is heavily complemented by actuators, drones and robots. The real-time situational awareness of such active systems requires sensed data to be transmitted in the uplink to edge-cloud, processed and control instructions transmitted in the downlink. Since many of these applications will be mission critical, the most suitable connectivity family will be cellular due to the availability of licensed spectrum able to protect the offered communications service. However, while much focus in the past was on the uplink of machine-type communications (MTC), little attention has been paid to the end-to-end reliability, latency and energy consumption comprising both up and downlinks. To address this gap, in this paper we focus on the definition, design and analysis of machine-type multicast service (MtMS). We discuss the different procedures that need to be re-designed for MtMS and we derive the most appropriate design drivers by analyzing different performance indicators such as scalability, reliability, latency and energy consumption. We also discuss the open issues to be considered in future research aimed at enhancing the capabilities of MtMS to support a wide variety of 5G IoT use cases.