I think that the best expressions of these three views are not actually in conflict. Indeed, I think the consequentialist does well to accept versions of all three.
The best form of scalar consequentialism merely holds that deontic concepts (e.g. right or ought) are not normatively fundamental -- we could fully account for the normative landscape without appeal to ought-claims, much as we (or idealized microphysicists) could fully account for the physical landscape without appeal to macro-object talk. At bottom, normative reasons are all that there are, and these reasons come in degrees (of normative strength or weightiness). But there's nothing stopping us from talking about non-fundamental "ought"-claims, just as scientific reductionists can happily talk about non-fundamental physical entities like tables and chairs.
I think that there are two deontic concepts in particular that are worth adding to our conceptual repertoire: what I call the "ought of most reason" and the "ought of minimal decency" (or hypothetical blamelessness). As you might have guessed, I see maximizing and satisficing consequentialisms (respectively) as providing the best accounts of the substantive moral contents of these two distinct kinds of oughts.
Maximizing act consequentialism provides, I think, a very plausible account of what we have most reason to do. One might reasonably worry that this standard is too demanding to correspond to the ordinary notion of moral obligation, but such worries are side-stepped entirely if we present maximizing consequentialism merely as an account of what we have most reason to do, and not as taking any stand on whether this ideal action is obligatory or supererogatory.
Something closer to the everyday concepts of moral obligation and permissibility may be captured by consideration of what acts are compatible with acting blamelessly, possessing a good will, or showing an adequate degree of concern for others (adequate, that is, to avoid warranting others' resentment). This presupposes that there are genuine rational norms governing emotions and reactive attitudes, but that seems right to me. (Importantly, a consequentialist about action can comfortably grant the existence of such non-consequentialist affective norms just as they may comfortably grant that there are non-consequentialist epistemic norms governing the rationality of belief. Consequentialism is only plausible as an account of the norms governing action, not other judgment-sensitive attitudes.)
Anyway, given such norms governing warranted resentment, we can construct derivative norms that apply to action, and these correspond pretty well, I think, to the idea of an "ought of minimal decency". And this is precisely the sort of less-demanding deontic concept to which satisficing consequentialism is addressed.
See my satisficing paper for a more fleshed out version of this argument. The paper is focused on developing a more plausible form of satisficing than has previously been offered (avoiding standard counterexamples, etc.). But one interesting upshot of the paper is the point that I've tried to bring out here: that one can comfortably accept forms of scalar, maximizing, and satisficing views simultaneously.