As a retired HW design engineer, married to an SW engineer, and the son of artists I have seen and found that you try, in the lab, many things before you find what works well.
What I found, especially as design engineer, is that the majority typically doesn't know what will work best; they just think the proper thing to do is obvious. Obvious solutions are seldom the best one, all things considered. With direct democracy the group will more often support the obvious answer not one of the better answers. The majority can take a look at a proposal by those they trust and those that have done their 'lab' work and can give a clear result and process that will work to get to the desired result.
I could have worked on this post some more, but I don't have time. So it's not the best explanation.