Much of this research is general and not Arts & Humanities focused. Basically, there is no conclusive answer, but the more ways you can get people to connect with your research the better. Time wise you will have to make a call about what methods you consider effective.
Niyazov et al. (2016) found that articles reached from Academia.edu gained more citations than those posted online on other platforms like personal webpages or the homepage of an academic institute
Thelwall & Kousha (2014) used bibliometric matching to show that popularity on academia does not equate to higher academic performance maybe “senior and active academics [do not] consider it necessary to spend time listing their publications in Academia.edu.” They are getting their visibility through traditional publishing.
Li & Gillet (2013) have shown that broadly more senior academics have more papers on ASNs and thus are more influential. This is calcifying the idea of research impact to traditional publishing outputs and they have argued for a new social measure of judging impact
ResearchGate and Academia have their own internal metrics. The RG score, and the top scholar score. There is necessarily a commercial interest to them in getting these adopted more widely within the academy. Quantative scores have always been a suspect way of judging quality, especially in the humanities