"Optimally," one would position each speaker at its "optimal" location (quotation marks because optimality will always be a compromise of one sort or another), and have only that speaker in the room. Then live with that speaker for a week or more before switching. BUT: this precludes any rapid A/B or A/B/X testing.
What I've done (and I've done this a lot; it's fun!) is to position two sets speakers of similar type (and so, similar placement needs) next to one another, and switch back and forth by switching the cables; if I could do this with a remote from the sweet spot, I would. This gives me a sense of which speaker I prefer. THEN I set that speaker up "optimally" for a longer listen. If I'm still uncertain of my preference, I do the same with the competitor.