After they've listened though, the devs should make their own decision, of course.
People don't know what they want. I have two personal examples if you (all) don't just think "tl;dr" to yourself:
1. My grandfather has worked in construction for I don't know how many years as a contractor. By the numbers he mostly build spec houses, or houses where there is no buyer before construction begins. The other 5% of what he builds is custom homes. In recent years, he has almost completely stopped building customs unless he knows the person he is building it for. The customers almost always prove difficult to keep in contact with, change their minds several times after it is too late, and make bad decisions when they do step in.
a] The first bit of that explains why he only builds homes for people he already knows. This guarantees that they understand he is good at what he does. They tell him what they're looking for, he makes the plans, then they approve and don't step in unless there is an actual issue. This makes everything work more smoothly because it's just in essence a more expensive spec house with a buyer lined up.
b] I myself have witnessed customers change their minds. First they want the wall painted one color, then "oh, I didn't realize it was going to look like that, can you change it back?" It's best to cut the buyer out of the loop for their own good. All they do is raise the price, tie up the painter/floor sander from working on other houses, and give the contractor nightmares. When you have two painters for sixteen houses, repainting rooms twice grinds the entire cycle to a halt. (All at different stages)
c] Finally, the buyers typically make bad decisions when they do make them. One person bought a parcel of land from my grandfather and wanted to build his own house on it. He started with a massive two story garage that is bigger than most houses. He never filled it more than halfway, and it was hideous when the house was finally built and attached as an addition to his garage. As I always said, it's a garage with a house attached. Indeed, when the time came for the man to sell his garage, no one would buy it (During the housing market boom, not rock bottom).
Custom houses don't sell because they're too tailored to a specific person/group. Think of these forums as the custom home buyers. We all want these different ideas placed into our game, we can't even all agree on the same thing. 2. I have worked in television production. One of the quickest lessons I learned was that customers don't know what they want. Unlike construction, we have to create custom videos for the clients. You have to listen to your client, but you must be grounded and know what is and is not feasible. If you overstep your abilities then you end up taking longer and making something with bad quality. The best thing to do is design what you think the audience will like and sell it well to the client. If you don't know what you're doing, then you shouldn't be a television producer; if the client knows better, then they should be the ones making the product.
Oblivion certainly shows signs of overstepping boundaries, so I am agreeing with the OP in that we need to not be so custom and exotic in our suggestions. Grounded ideas are the best. Think of ways to spice up the game without being ridiculous (Pets vs hardcoe Physics).