A sometimes unrecognized problem is that you cannot determine product/market fit without actually building a real product. See my post The Minimum Viable Product.
Several people have posted to Medium discussing their experience polling potential users about a product that they considered building. They found that people are quick to respond positively, but do not necessarily have the same response once the are presented with the actual product. In many cases surveys of potential customers are illusionary.
Of course it all depends on what you’re building. Some things can be hacked up, tested and then rebuilt if they show promise. The problem with the dogma of the MVP is that it does not fit all cases.
In the case of nderground, we are building a social network. The bar for a viable social network is high. If the initial product does not have enough features and it not reliable enough, people are not going to use it. Even if the idea is promising. We don’t have to have all of the features that we plan to build, but there has to be enough there to attract and keep users.
In my judgement a complex product means that you have to build something real, with a solid foundation. That means building something that scales, because anything that doesn’t scale is not going to be a product going forward.
With a complicated software system you are, necessarily, building with the assumption that you’re going to succeed.
I find it interesting that both Charles Frank and I had some of the same impressions from your article. Perhaps it’s our sensitivity from being experienced engineers in an industry that discriminates against experience. Or perhaps there really a subtext in what you wrote. Food for thought, perhaps.