It really depends on which experiment you look at, but the one that I know best which was the one done in Chile managed to capture a fair few demographic catego

It really depends on which experiment you look at, but the one that I know best which was the one done in Chile managed to capture a fair few demographic categories. You can be as granular as you want it really depends if you have the means to do it, and if you think about this from the state-perspective, you would most definitely be able to capture as much granularity as you wanted. Generally, I would say though, capturing major demographic categories (age, gender, sexuality, socioeconomic status) tends to be enough to produce very good outcomes. Specially when it comes to socioeconomic status which is definitely one of the biggest concerns with oligarchic politics. I mean, think about how much money and time it actually costs to run for congressman or senator.