Cloud services are changing the software development context and are expected to increase dramatically in the forthcoming years. Within the cloud context, platform-as-a-service tools emerge as an important segment with an expected yearly growth between 25 to 50 % in the next decade. These tools enable businesses to design and deploy new applications easily, thereby reducing operational expenses and time to market. This is increasingly important due to the lack of professional developers and it also raises a long standing issue in computer-aided software engineering: the need for easy to learn (low-threshold), functional (high-ceiling) tools enabling non-experts to create and adapt new cloud services. Despite their importance and impact, no research to date addressed the measurement of tools’ ceiling and threshold. In this paper we describe a first attempt to advance the state of the art in this area through an in-depth usability study of platform-as-a-service tools in terms of their threshold (learnability) and ceiling (functionality). The measured learnability issues evidenced a strong positive correlation with usability defects and a weaker correlation with performance. Remarkably, the fastest and easiest to use and learn tool falls into the low-threshold/low-ceiling pattern.