I like the membrane tank analogy of a capacitor as well. It also brings up a common misconception of capacitors, when we say that they "store charge".
When the membrane is in the unstressed neutral position, is the tank completely "diswatered", just as a capacitor would be "discharged"? It contains no water at all...right?
When the membrane tank is in an extreme position, does this "heavily watered" membrane tank contain more water than before?
Interestingly enough, the total amount of water in the membrane tank is the same, no matter what the position of the membrane. So it isn't storing any additional water by pumping it from one side to the other. It stores the same amount of water no matter what, and what it really stores, is energy in the elastic deformation of the membrane.
Similarly for a capacitor, it always has the same net charge (i.e. zero), and it always has the same population of electrons. It isn't storing any charge in the unit as a whole, when one plate has an electron surplus and the other an electron deficiency. So it is really storing energy in the electric field between the plates, due to the non-neutral distribution of charge, when the capacitor is energized ("charged").