Even 1 miligram is probably a lot of DNA, but I wonder how to know if the place of sampling is important
Sampling and storage methods can significantly change the bacterial composition of an environmental sample (in this case, poop). The exact protocols will depend on the aims of the study. The gut microbiome is a gradient and very dynamic. Different parts of the gut will have different bacterial compositions. Some people might prefer to get a locally accurate sample from a biopsy of the intestine, but you won't manage to recruit many participants. Other studies may prefer to use faecal samples as a proxy for overall gut state, which lets you recruit more people. Some protocols may homogenise (blend) the poop before sampling, others might not. Here's a nice review:
You can go as low as 10ng depending on the library you use (or so the vendors say), but I'm not sure it's the case for these specific applications (my experience is with other, equally difficult samples, but from a single source).