This paper measures the performance in terms of costs of Swiss drinking water utilities accounting for environmental factors. We estimate a translog stochastic variable cost frontier using two different techniques on an unbalanced panel of 141 water distribution utilities over the years 2002-2009, for a total of 745 observations. Results show that exogenous factors have an impact on variable cost. More precisely, we find that the share of pumped over total extracted water, population density, altitude and meteorological factors (maximum 30 days temperature and extreme precipitation events) have a significant impact on variable cost. Likelihood ratio tests emphasize the importance to include observed heterogeneity in the estimations. Efficiency rankings provided by models accounting for exogenous factors and their counterparts without them are however relatively similar. On the contrary, the efficiency ranks differ strongly between alternative estimation techniques. In assessing the economic performance of utilities, the most important choice thus seems to be about the way unobserved heterogeneity is treated.