The great number of e-business websites existing over internet make the customers undecided on which one is the more appropriate for their purchase. Thus recommendation websites have appeared. They propose an aggregation of the customer's evaluations critics on the e-business websites according to several criteria such as prices and conditions offers, after sales service and payment security, etc. In fact a partial score relative to an evaluation criterion is associated to each critic. It's either a precise numerical value or a purely qualitative value. Then, an average upon the global scores of the customers is used to evaluate the average e-business global score. This drastic reduction to average scores prevents the users from really appreciating the uncertainty in an e business website evaluation. Indeed, it does not give any idea about the variability or controversial dimension of the customers' opinions. All these aspects gathered bias the final decision. One way to overcome this problem is to consider a possibility theory based uncertainty representation and to construct possibility distributions from the customer's evaluations according to each criterion. Possibility distribution representations of the customer's evaluation allow thus to take into account the whole information in the decision making process. We address the issue of how to interpret the aggregated uncertainty representation in the global evaluation. Indeed, the uncertainty analysis of the system was conducted for the purpose of aiding in a decision. Thus, concise easily interpretable pieces of information, such as variability or controversial aspect of the customer's opinions, are useful. In this view, we propose to describe possibility distributions by indicators of location, ignorance, divergence, and to determine the contribution of each partial indicators to the global indicators.