One way to qauntify this form of total effeciency would be calulate speed/power expended per second or even, as used in reference to cars, distance traveled per unit of energy expended. The problem here is of course that extarnal factors will greatly influence the readings. You would need to take both measurements in an abstract world with no wind or hills, or a constant grade, and for distance/energy you would need a constant control speed.
Although hard to quantify, this concept is a useful one in comparing cyclists, expecially in reference to TT's. For example, riders like Ekimov who were lighter than Lance could not hold him on the hills- they had lower power to weight ratios. They could do okay against him in TT's however because their total efficiency was better, due to better tt positions and narrower bodies lowering aerodynamic drag. (of course, here I have used only one part, speed/power, not speed/power expended.) To get the latter you could take for example Vo2 (thres.) / average speed for a flat 40k TT course, no wind.