Variability of traf?c conditions is a well known fact. Traditionally, vehicles have been developed to cope with many very different conditions, both in terms of environment and traf?c, albeit at the price of, among other, reduced ef?ciency under many conditions. The progressive increase of on board computational power, sensors and other available information offers new possibilities, in particular, besides achieving their main function, ADAS can be tailored to achieve additional bene?ts, e.g. reduce fuel consumption or improve driving comfort. However, we may expect that the levels of improvement strongly depend on the speci?c traf?c conditions. To check this hypothesis, in this paper we use measured data to build up clusters of traf?c situations, and then analyze by simulation the achievable improvements of fuel consumption vs. comfort at the example of a speci?c ADAS, both in the case of averages and of single vehicles from the clusters. As the examples con?rm, the trade-off is indeed context dependent, and control tuning should be adapted accordingly.