I think algebraic merkle trees (curve trees) are a natural fit because you get fast verification (10-50ms without batching) and fast proof (relatively fast; a second or two), because of not having to put a cryptographic hash function through arithmetization. And you still get the same ability to embed logical conditions on what is in the leaves of the tree, using standard ZK proving systems like bulletproofs (e.g. I did "proof of reserves" in aut-ct). Maybe worth looking at Luke Parker's work on FCMP for how much this can scale. I described aut-ct more wordily here if it's interesting: https://delvingbitcoin.org/t/anonymous-usage-tokens-from-curve-trees-or-autct/862 .
I also think reputation is a false trail but, meh, that's just a vague opinion.