Another thing I observed was building something (at least) twice is a valid development process.
1. The first time, the client hammers out their (often changing) requirements and you get a pretty good understanding of the problem that actually needs to be solved. So the main objective of the first build is actually just gathering the requirements. You also uncover issues that may have been missed just "thinking" about the requirements.
2. The second time almost never takes as long to develop, so the development time to build that first version is a good hard upper-limit. It usually takes significantly less time because development of the second version usually involves copy-pasting a lot of the original code and just refining it.
1. The first time, the client hammers out their (often changing) requirements and you get a pretty good understanding of the problem that actually needs to be solved. So the main objective of the first build is actually just gathering the requirements. You also uncover issues that may have been missed just "thinking" about the requirements.
2. The second time almost never takes as long to develop, so the development time to build that first version is a good hard upper-limit. It usually takes significantly less time because development of the second version usually involves copy-pasting a lot of the original code and just refining it.