Table 7 shows the annual metric tons of CO2 emissions from the 28 sample hotels. In order to find the energy use indexes responsible for the CO2 emissions, correlation analysis was carried out and the result is displayed in Table 8. Results show that the energy consumption per unit guestroom and energy consumption per unit equivalent guestroom have significant correlations with CO2 emission.Statistical options for benchmarking annual energy consumption per unit guestroom and annual energy consumption per unit equivalent guestroom are shown in Table 9. From Figs. 2 and 3, mean is a suitable benchmark for energy consumption per unit equivalent guestroom but it cannot be used for energy consumption per unit guestroom because of the high CO2 emission associated with its value. 75th percentile, rather than 60th percentile, is a better benchmark for maximum normalized energy consumption because increase in energy consumption per unit guestroom and energy consumption per unit equivalent guestroom after the 60th percentile still present opportunities for lower CO2 emission. However, increase of normalized energy consumption after the 75th percentile will lead to extreme CO2 emission.