The complaint that "one can't define it" is not an argument against the existence of a phenomenon. Philosophy IS the collection of subjects that we do not understand well enough to be their own specialty fields, hence most terms, subjects, and fields in philosophy intrinsically are difficult to arrive at full consensus on definitions and methodologies of investigation.
Consciousness is an observed phenomena. It includes an essential component of internal awareness. As experienced in people, it has a variety of features and functions. These include CONTRADICTORY features and functions, such as an integrative feature, and an experience of unity of consciousness, yet demonstrated discontinuities in experience, and multiplicity of experience and recollection. The predictive function using world-modeling and hypothesis testing in a virtual space, and the analytical processing function where we linearly think through a problem stepwise, both COULD be done unconsciously (we have created unconscious machines to do both), but happen to be conscious in humans.
The observation, compilation, and characterization of our consciousness is not something that lends itself to a "definitional" approach that is characteristic of analytic philosophy. Instead, it is a field that lends itself to an empirical and pragmatic approach -- basically following the first few stages of the scientific process, which involve exploration and characterization of a field of study, followed by speculation and tuning of postulations about that field. A number of researchers have attempted the next steps, which are more formal hypothesis forming and testing, which needs to happen for a field to separate off out of philosophy, but so far without any consensus on their success.
The evolutionary role of consciousness seems to be to make free will choices. This is not refuted by the observation that aspects of our brains make those choices before we are aware of them. It is useful to spell out the assumptions behind the invalid claim that they show prior brain determinism over mind. Those assumptions include the belief that there is ONLY ONE such prior decision! Daniel Dennett has proposed a "multiple drafts" model of brain function, in which our brains perform multiple parallel hypothesis models, EACH of which we start preparing to implement, and there is a fair amount of evidence for this model's validity. ONE draft then comes to dominate over the others, and harnesses the entire brain to implement its course of action. The timing of a "first preparation" is what the brain readiness studies looked at, and in a multiple drafts model, this MUST occur before the collapse to one implementation. Dennett himself denies the existence of a decider, but use of his model, and the association of the collapse to one implementation as the point of "conscious decision", is entirely consistent with the lab data, and consciousness decision being causal.
Note, the modular/tuned evolutionary nature of the mind is only compatible with the mind being causal. This is the argument against epiphenomenalism made by William James, and it is equally effective against the Identity Theory more recently favored by materialists.
This problem, that the mind is tuned to be causal, but if matter were causal on its own, then evolution would not have bothered with a mind (and if it did, our mind activities would be unrelated to what we actually do), is the core of the "Hard Problem of Consciousness". If reductive materialism or functionalism were true, then we SHOULD NOT have an evolved/tuned consciousness. You yourself make this point -- in a materialist reductionist worldview, the mind doesn't matter. Yet, evolutionarily, it clearly does.
Some determined materialists have rejected the existence of consciousness despite the observations. This POV is called Delusionism, and its leading advocate is Daniel Dennett. Dennett is a subtle and cagey philosopher, who often seeks to persuade indirectly rather than overtly, so the rationale for delusionism is not clearly articulated by him. The clearest justification for it is found in Susan Blackmore's A Very Short Introduction to Consciousness. She spells out a belief that materialism is absolutely demonstrated by physics, then identifies a series of laboratory tests that show that consciousness cannot be a purely materialist product of our brains. She then concludes, since consciousness cannot be material, but if it exists it must be, that consciousness cannot exist.
Blackmore's rejection of dualism does not actually cite the physics she asserts to be constraining. In his essay for The Myth of the Afterlife, Augustine cites two physicists who refute the conservation of energy argument made by most materialist philosophers against dualism. Papineau's The Rise of Physicalism essay notes that dualism is not actually refuted by any aspect of science, it was instead widely abandoned when biochemical reductionism proved to be so much more useful in characterizing cellular life. The limits to reductionism that have lead science to abandon the Unity of Science programme, bring even Papineau's "utility" rationale into question. There is also at least one alternatives that Blackmore does not consider, that of causally independent emergence, which is what Karl Popper advocated for (See The Self and Its Brain).