Simplicity and validity are recognized as important attributes of an effective simulation model. However, some disagreement exists both on the definitions of these concepts and on quantitative measures for them. This paper introduces Enrichment Level (EL) for (simulation) models. The EL is a quantitative measure to compare the effectiveness of alternative models when used in a particular project. While there is no conceptual restriction on expanding the EL, it currently captures the bias, speed, variance, and scope of a model as the main contributing enrichment factors. This paper explores these factors and suggests a utility-based method to combine them according to the relative importance of each enrichment factor. Model effectiveness measures currently used in academic and industrial studies are discussed and compared to the proposed EL. In order to show the usability of EL as a model comparison tool, we study an example of a queue with multi-failure mode servers and apply our method to facilitate decisions about choosing the most suitable model among a set of three.