The Internet of Things (IoT) is a key enabler for modern digital societies by providing sensor information for smart and sustainable applications. In this work, a performance evaluation of cellular IoT networks is presented. Using analytical models, simulative approaches and empirical measurements, all facets of challenging IoT scenarios are quantitatively investigated to identify appropriate solutions and optimal system configurations. Based on extensive coverage measurements and simulations in sub-1 GHz frequency bands, the performance of the technologies NB-IoT and eMTC in terms of the complex interdependencies between battery life, signal range and spectral efficiency. In terms of massive scalability, an extension of a network simulator was realized and used to identify bottlenecks and optimize parameters in current public and future private NB-IoT configurations. A novel 5G energy planner supports IoT application developers by translating results of the complex system models into easy-to-understand decision criteria for application design choices. This work therefore not only provides a detailed quantitative assessment of the performance of cellular IoT technologies for wireless communication experts, but also contributes to a better understanding of the impact of different solutions in the IoT application developer community.