It depends on if these are paying users or not. I'm assuming yes so the answer is simple, the ones that will pay for it. I heard a great analogy recently that

It depends on if these are paying users or not. I'm assuming yes so the answer is simple, the ones that will pay for it. I heard a great analogy recently that a startup is basically like consulting for a very specific use case of customers in the early days. One of the trickiest thing in building a product users love is figuring out what feedback is fluff and what is real. The best way to determine this is to first see if there are common patterns in the feedback your are getting. If yes, that's a good sign. Next is to see if you can get a few of these users to pay for it. If they'll pay for it before it's even built, you know it's a real problem. If it's something they will commit to pay for (on paper) once it's built, that's still good but be more cautious. If you get excuses or any other non-comital response, then it's more than likely not something worth building right now. One cautionary note here is to be careful not to go down a rabbit hole for one customer. You don't want to start building features that are just applicable to one specific customer early on unless they are paying you a significant amount of money. The ideal scenario for building a software product is to leverage feature requests that multiple users/customer have toward the long-term objectives of the product.