Cloud computing has gained popularity over the last years, causing a significant increase of available cloud offerings among providers. Therefore, this wide spectrum of options has led to an increment of possibilities for distributing applications in the cloud, by means of selecting specialized services to host each application component. Nevertheless, it also implies the need of finding the optimal solution depending on its purpose, usually based on future economical profitability. Nowadays, instead of considering an application as a whole when deploying it in the cloud, e.g. deploying whole application stack in a virtual machine, investigations focus on how to distribute the application components in heterogeneous cloud environments. Consequently, users have an even higher range of options and should carefully choose good decision criterion, going further than only considering the direct cost for the needed cloud instances. Some challenges are deriving a revenue model - as they tend to be application specific - and customizing the evaluation of different migration configurations of a real application with authentic data metrics. In this sense, this document uses utility analysis as it includes a non-directly countable element, preferences, and allows basing the decision on a trade-off taking into account other aspects which have an influence on the final performance such as users satisfaction or cloud instance availability under different deployment topologies. Therefore, the evaluation and comparison of different selected cloud offerings is possible and helps throughout the decision. This thesis presents an overview of state-of-the-art revenue models used nowadays on web applications and afterwards specifies the study and aims to apply the utility concept to evaluate a current application, MediaWiki, based on real data. Results show that this approach is more complex and differs from the one considering only the monetary expenses, pursuing a better balance between the possible business-technology conflict.