The first lab I ever worked in at NHGRI had a saying: "We do the 're' in research".

I can understand how when you are bringing together unified mathematical ideas based on centuries of other people's data, if you have the right insight you can figure some major things out quickly. But when dealing with complicated real-world systems rather than abstract logical structures you have to keep pumping. Sometimes you get lucky or you have the correct thought as to what is interesting and where to keep pushing, sometimes you need to keep pushing and pushing away.