That sounds very much like my criticism of today's software development habits for research settings (
https://doi.org/10.1145/3689492.3689808). In research, the goal is not a product but understanding of the problem space. What you say, in summary, is that even if your goal is a product, you must first understand the problem space.
What was new to me is the enormous list of techniques proposed, but apparently not so much used, for applying this principle in engineering practice.