Actual real world community management:
1. Idea is economically useless - serious investment vs low return. Or a direct subversion of revenue:
You can never tell the user that their obvious improvement is totally correct but your design demand inconvenience so relief can be monetized.
For any other tech heavy stuff - employing a two step approach is preferable. At step one you let them know you are going to pass it on (flag your milestone, and keep track to come back later). At point two you say you've checked with dev team and they say it takes too much time for too little benefit. If you are well embedded in the work-flow you can often make these decisions yourself without asking coders or designers. If your dev team actually wants solid feedback - you put these in to well comprised lists so devs can actually take a look at them in an organized fashion.
The only thing you have to do is make sure your assessment is solid. If you ever reject a SIMPLE fix with this statement for any reason, there is going to be a massive surge in credibility, as someone technical can say something like "Get serious, fixing a minor html bug on that page takes 5 minutes of a 10 year old - here - I'll give you the code for free".
2. Post is in unintelligible:
"Can you please describe what is it that you suggest EXACTLY and more clearly, because I'm confused."
3. The idea is outside the goal:
"We have have considered this option, but for now it is now what we are looking to do" often its enough, if the honest reason why is acceptable give it. Otherwise you may better lie and say its possible to happen in long term future, but its very low on the priorities list.
4. Idea is a luxury:
Something kind of cool, but is really an EXTRA BONUS not a NEED.
I usually maintain a list of these, where devs can handpick them when they want to. Again i say to users this is added to a list of low priority possible improvements.
What you can and cannot say is heavily based on your general disposition and behavior.
The user-base I work with trusts me because I have indeed officially credited people for suggestions that were implemented. I also have taken to account circumstances and I have given very detailed explanations where they weren't necessary.
If the general perception of a/the community manager is that he is paying attention. It is very bankable in the long run. You can close cases/threads/posts/tweets with a set of 10 template answers - and even if you get questioned - the regulars in the user-base are going to verify your credibility and that what you are saying holds weight.
I view it more or less like a currency, you earn some when possible (by going out of your way to give more than the average support hypocritical bland answer) - and then you spend some when brushing off people you don't want to deal with without having to explain yourself.
If you've always copy pasted templates people view you more or less as a drone trying to get rid of as many requests as possible - without ever trying to communicate with them on a person to person basis. You are never going to have a positive environment where people feel their expression is just wasted in a forum/whatever platform. And they don't need to know you. They can just list a few threads and see the proof that you indeed have taken another-random someone (just like them) and have bothered to engage with them thoroughly. Why didn't you enter extensive dialog with me? - Maybe you were busy or indeed my post has lees merit than I think. Its obvious though that you CAN engage and that you HAVE done so in the environment.
Following the all famous corporate official like behavior pattern is a slap in the face of anyone who is emotional or passionate about a thing. And its usually these people that bother diverting time from their lives to explain to you why and how you have to improve your product that you are profiting from.
And despite what many people think, being human even miss-spelled and with grammar errors HELPS. Better have a genuine communication with a few misspells, than a packet of honed in "perfect" cold-read copy-pastes. If your support environment is limited and you re-use templates, it comes a time people assume an AI is handling basic communication. And whats the point if having people to handle people if they are going to be perceived as machines.
Obviously none of this is possible if management never allows you to be a person on-line.
This mistake is made very often with games. Players come to have fun in games. When your rep handles them like a pro lawyer explains the 367 excerpt of a company merger contract to a committee. You instantly trigger the thought "Oh damn it, again its another suit that wants to drain my money and only has a forum/etc because - everybody else has or its traditional - they don't care"...
It is regular to run a company on that old tradition. But this is the lowest benefit factor you gain from a community.
There is no conclusive study on the greater benefit of community management. However a successfully positive community is easily felt anywhere it exists. And any fool in business can tell you a satisfied customer hold great referral value that is often immeasurable. Thats why some smarter companies out there handle support publicly. They want to world to know how their employees handle their other customers - because they've made an effort in there.