To reduce global greenhouse gas emissions associated with building energy use, owners of large building portfolios such as university campuses frequently rely on building energy models (BEM) to better understand potential costs and benefits of retrofits. Model development workflows that are designed for individual buildings require a level of effort that would be time and cost prohibitive to apply to such campuses which often include hundreds of diverse-use buildings. While smaller campuses can effectively utilize the traditional BEM approach to study retrofit scenarios, this option is therefore not feasible for larger campuses. Large universities have instead utilized a combination of statistical and spreadsheet models which may not fully capture the unique architectural features, programmatic requirements, and systems configurations of individual campus buildings. With the goal of overcoming these limitations, two separate urban energy models, that employed considerably different methodologies, were developed for the campus of the Massachusetts Institute of Technology to evaluate future energy scenarios with an appreciably smaller effort vis-a-vis developing building-by-building energy models. This study reviews these two models with regards to their setup and calibration effort, ability to model individual building energy use, and the accuracy in predicted savings from implementing a variety of retrofitting measures. The paper identifies both models' strengths and limitations and suggests best practice procedures for administrators of other campuses interested in undergoing a similar exercise. (C) 2018 Elsevier B.V. All rights reserved.