You make a fair point here. You are getting at ship fast, fail faster and iterate.
This would not be the case with very wicked, high-level or technology-driven problems. For example, we worked with Intel in grad school to find potential ways to use their RealSense cameras — they are basically depth sensing cameras. We have the technology, but we were asked to find what to do with it. Here the problem is extremely broad. We had teams that used it for music, living spaces at home, to even open environments. Such a technology-driven design problem would not be here in the first place if there was more research on what to invest time/money on. I am not saying this technology is going to be useless, definitely there are multiple avenues this can be used. I am trying to iterate on Fabricio’s point of taking a step back and looking at it from a broader lens helps.
Even Facebook went through intense research for implementing their reaction patterns. When you are designing for such a scale, it is easy to forget or miss sections of people or different types of users. Of course you can get back and iterate, but getting something substantially good as opposed to something that just works is the difference we are talking about here.
Having said that, I completely agree with phased research periods where we don’t spend way too much time just researching and having nothing out there. It is a fine balance that is quite hard to achieve.