I've worked in multiple startups/bootstrapped enterprises that started as consultancies to pay for the development of internal tools with the ultimate goal of l

I've worked in multiple startups/bootstrapped enterprises that started as consultancies to pay for the development of internal tools with the ultimate goal of licensing the tools. It doesn't always work but its the least bad validation strategy I've seen. I find 1) and 2) of your "bad" items to be more nuanced. Mostly because outside of software development most industries do not have the capacity to build a solution outside their particular flavor of a problem, and they don't do it internally. They hire consultants to do it, and there are enormous inefficiencies that prevent their solution from turning into a product.