Emergency department (ED) overcrowding has been recognized as a serious concern in hospitals nationwide. In response, many hospitals have implemented various improvement initiatives, but not all hospitals have found great success. This study aimed to develop a data-driven framework for benchmarking efficient EDs and adopting their best practices. First, a data envelopment analysis (DEA) was used to identify ED frontiers that have achieved operational efficiencies. Using the Emergency Department Benchmarking Alliance database for 2012, we divided 449 EDs into six groups and evaluated the efficiency of each ED in the groups. Inputs include number of ED beds, clinical staffing working hours, and non-clinical staffing working hours in the ED. Outputs include the number of patient visits per day, average length of stay, and the rate of leaving without being treated. Using the efficiency rankings, logistic regression was performed to identify which features of EDs contributed to their becoming efficient frontiers or inefficient units. The results indicated that the proportion of admitted patients through the ED, the intake model with a mid-level provider, a fast-track area, and a patient volume had a significant impact on efficiency of the EDs. The identification of successful EDs and their profiles may provide comparable ED benchmarking and the proliferation of best practices.