As a PM focusing mostly on new product development, my mantra to engineering was "Do the easy, cheap stuff first." The corollary was, "If we are very lucky, we will live in regret because the product has taken off but the architecture is now obviously wrong." This was understandably frustrating to many engineers because they wanted me to have a certainty about the correctness of what we were building that I just never had. Every good idea I had for product emerged over time, often years and countless false steps. I could never figure out how to avoid the false steps so the goal became to keep cost and effort down along the way.