An organization that would engage in a costly New Product Development process "in the vacuum", devoided of any input from potential customers, users and consumers should be awarded the
Darwin award, in the corporate category.
Many market needs are obvious based on documented needs and/or common sense. This is the "if we build it, they will come" marketing philosophy.
Other market needs are met in backdoor fashion when an inventor creates something that he or she uniquely realizes how to do, often purely because they enjoy creating things and overcoming technical challenges, and it turns out that he or she isn't the only person that wants the invention. This is the intuitive path.
Some examples:
1. A pharma company, as part of its broad mega$$$ search for biomedical action of complex natural organics, discovers a substance that achieves an improved level of effectiveness in treatment of a known disease...so it productizes it. There's no need to consult with potential users and disease sufferers, at the point of the initial productization decision, as to whether they're interested in a more effective treatment option than they currently have.
2. A teenager codes up a functionality that he would like to have exist, because it doesn't and he can. It turns out that a bunch of his buddies like it too. Then it turns out that 100 million other people want it as well. That's basically the story of DOS, Windows, Lotus 1-2-3, online "bulletin board" technology, Facebook, Twitter, Napster, etc.
3. The first personal computers were all hobby or hobby-business efforts that turned into businesses, and eventually a rather large industry. Often design was driven, not by what customers would want, but what parts the designers could afford to buy, or existed to be bought, or could finagle someone into co-developing for them.
4. Lockheed is one example of a rather large company that for decades maintained an expensive sub-operation (in their case, the so-called Skunk Works) tasked specifically with building things that they could figure out how to do, and would have neat new capabilities. No customers involved until they had something that flew and was "interesting".
None of these is the norm, of course--the pharma example is less common than it was in the '70s, say, because a lot of the world has been explored and because molecular synthesation has proceeded to the point that design now more frequently is marketing driven, i.e. model and test all the possible molecules that have these characteristics, similarities and interactions, and meet these criteria--but they've all been historically important.