A sociobiological explanation -- naturally a speculative just so story like almost all evo bio stories -- is that non-alpha males have partially genetically based characteristics that it would bne useful to the survival of the group, thus of the alpha males themselves, and maybe the females, to pass along, so there may be a partially genetically based dispositions to adopt arrangements where the alpha males do not monopolize all or almost all the females.
The disposition must be pretty weak, since polygamous groups and societies are not uncommon, however.
[WS:] There is a simpler alternative explanation: what matters in evolution/natural selection is one and only thing - the chance of survival. The traits that increase that chance get transmitted to further generations, the ones that decrease that chance eventually become extinct.
It can be argued that large polygamous social units in which the alpha male dominates all or most females have a lower chance of survival than social units in which females are "spread out" so to speak among multiple males. The reason is quite similar to that why people spread out their investments - the failure of one does not translate into a total loss, only a partial one. In a large polygamous unit, the failure of one male can decrease the chance of survival of the entire offspring pool of that unit. In a multiple monogamous units society, such a failure would jeopardize but a small fraction of that offspring pool.
This is sheer law of probability that has nothing to do with prowess, skills, strength, sex appeal or for that matter "the survival of the fittest."
Wojtek