There is a significant potential for optimizing pump systems currently in use in groundwater wells. This potential lies in: (i) the improvement in pump technology, which can yield up to ~5% more efficiency, (ii) the improvement in motor technology, which can yield up to ~3% more efficiency, with further improvements if innovations from aboveground motors are adapted, (iii) the improvement in performance adaptability, which can be very efficient in some cases (~10-50%), but also counterproductive if not adapted to current situation (0% or even efficiency loss), and sometimes not very flexible (impeller trimming); (iv) the improvement of the system maintenance and management which may yield up to ~20% more efficiency, and which, in general, has a shorter payback time than performance adaptability options.The improvement of equipments may induce only moderate additional costs if it is done at the time of scheduled new investments, after amortization of the equipment formerly in use. Unfortunately, these expected savings are influenced by uncertainties, which can be of the same order of magnitude as the savings themselves. For instance, the determination of the optimal operation point of a pump bears uncertainties between 1% and 4% and grows with pump rotation speed (Gülich 2010). Other considerable saving potentials lie within cleaning, maintenance and smart wellfield operation with short to moderate payback times (Table 6). These potentials are however very site-specific, and difficult to estimate on a general basis. Best practices for a “smart” pumping shall include choosing equipment that fits the actual requirements of the system, operating the pumps nearest of their Best Efficiency Point, and operating the motors in an energy-efficient load range. The most obvious energy savings are those associated with improvements in the efficiency of the motor and of the pump (Shiels 1998). Such gains are often worth the added capital expenditure – although often having moderate to long payback times. However, as underlined by (Kaya, Yagmur et al. 2008), that pumps have high efficiency alone is not enough for a pump system to work in maximum efficiency. An improvement of pump technology will yield, even optimistically seen, an efficiency improvement of up to 10%, which is the potential “theoretical limit” (EC 2003). For further improvements, it is necessary to consider solutions that go beyond the pump system, since maximizing efficiency depends not only on a good pump design, but also on a good system design. Even the most efficient pump in a system that has been wrongly designed is going to be inefficient. Moreover, an efficient pump in an inefficient well is pointless. Hence, a global approach of the groundwater abstraction system is required. The optimization potentials highly depend on the site characteristics themselves, on the local demand (what distribution of the demand? what load profile?), and on the operation and maintenance history (e.g., what is the cleaning frequency of the pipes, if any?). Finally, one should not forget the primary objective of water abstraction, which is satisfying a given water demand, thus, the safety of drinking water production prevails over energy efficiency.