The Internet of Things (IoT) enables connectivity between devices, thereby allowing them to interact with each other. A recurring problem is the emergence of siloed IoT platforms due to proprietary standards. Recently, the World Wide Web Consortium (W3C) proposed a human-readable and machine- understandable format called Thing Description (TD). It allows to uniformly describe device and service interfaces of different IoT standards with syntactic and semantic information, and hence enables semantic interoperability. However, describing sequential behavior of devices, which is essential for many cyber-physical systems, is not covered. In this paper, we propose a systematic way to describe such sequential behavior as an extension within TDs, thereby increasing their semantic expressiveness through possible, valid state transitions. This enables safe and desired operation of devices as well as scalability by modeling systems as sequential compositions of Things. We show in a case study that previously unmodelable behavior can now be expressed and the overall manual intervention requirements of state-of-the-art implementations can be significantly reduced.