The interest in summarizing complex and multidimensional phenomena such as development, sustainability, poverty, quality of life, and well-being for broader comprehension, even among non-experts, is far from waning. Composite indexes stand out as the most commonly utilized methodological tool for this purpose. Various composite indexes have been proposed in the literature, often driven by the need to streamline a particular phenomenon. Each index is built upon specific assumptions, and its robustness for a given application is often evaluated afterwards, through uncertainty and sensitivity analyses. The performance of an index relies on the nature and characteristics of the indicators it aggregates, as well as on the relationships among them. We address these questions by investigating, through a simulation study, two distinct approaches to construct a composite index. The objective is to identify elements that, considering the available data and their characteristics, enable the selection of the most appropriate synthesis rule whatever is the multidimensional phenomenon under investigation.