One of the easiest parts of designing a product, in my opinion, is dreaming up features and things that the user should be able to do. The harder part, the part that makes my brain part, is sitting there and wondering, "Does this feature really need to be here?" Whether you're moving that feature to another place, or whether that feature needs to be there at all, it's the part of design where there's no clearcut recipe - the art, if you will. Perhaps the entirety of the process is about making this moment of intuition work: the more deeply and fully you understand what you're designing for, the more intelligently you can make these decisions.
It's true that you're often iterating and perfecting as you go along, but this act of whether to include something or not is a bit like choosing a road - except you've got no map, just the intuition of what direction is "right." Sure, you might eventually be able to get back in another direction, but particularly in the world of social products, once you've established momentum in certain direction, it's difficult to really get back to that starting point. For example, I recently learned that MySpace was originally started by spammers - yes, that Tom originally started MySpace to bolster their flailing spam business. Sure, there were other people involved in MySpace's long-term development, and the founders developed a compassion and sense for what their users wanted, but you tell me: how far was MySpace able to get from the original intuition that informed it?
Another thing I've been thinking about recently is one of the things designers say about brainstorms: Crap in, Crap out. I'm a big, big believer in doing the research before designing a product. Before I even started thinking of potential designs for a product supporting communities, I dove into a few months of research. And not just competitive research, though I did do a bit of that as well (but actually, more towards the end of my process, after I'd already had the basics of the product designed), but research with users and communities. I used this research to explicitly think about what needs the product needed to solve and what it should be, but the effects are lasting beyond that: it's also developed a certain level of intuition for me. A certain sense for the patterns and behaviors, if not emotions and motivations, that a user might have in relation to a product. I'm not, by the by, saying that it's right, but the intuition is certainly there.
And I feel it most when I decide whether to remove a feature or let it stay. There's no autopilot for that moment - I actually have to stop, and force myself to really think about the question. What are the contexts the user might be using this in? What's the story as the user comes through? What situations do users want this feature in? What are situations the user doesn't want it? Is this the kind of interaction I want to promote? I stop and tell myself little stories, try to feel what a user might feel, even if for just thirty seconds. And that's the part that makes removing features hard: you really have to be thinking deeply about what you're designing, because it's that quality of thought that will reflect in your product.