:) The question was hypothetical. Besides being of interest to any small cash-strapped startup, your answers are being used in the construction of an AI system. Knowing the factors that are most important to a venture's contributors will help us determine an algorithm to value the contributions.
In one ongoing series of consumer products, my several-hundred-thousand-line codebase lasted for well over 10 years. There was no reason to mess with it. Yes, I thought parts were messy... but there was also great stuff in it that I'm very happy survived intact.
Then decades later, on an entirely different product, in a different business...
We had to migrate our product to a new OS platform, as well as needing to fulfill much wider requirements for our customers. The existing code just wouldn't allow this, even if it were stretched and stretched.
So I rewrote most of the codebase.
In some ways I was tortured knowing we had a functioning product that already handled things I was rewriting (even though it was on an obsolete platform). Especially since the rewriting process takes longer than zero.
On the other hand, the new set of code turned out smaller, simpler, more reliable, much more flexible, and is now working well as a foundation our for business.
Both sides of the argument make sense to me, I think it just depends on the situation.
I agree: overnight success may be programmed over a weekend, but it also has to be grown with care like a garden.
I've heard "buy some customers" a few times myself, and while it's possible for me to think like that, who as a customer would want to be treated that way? Not me. This article is an excellent, level-headed summary.
There's always a danger when building on top of a platform that's not designed as an "infrastructure" service. Even public utilities like power & water aren't 100% reliable.
When you pick a platform to reach your customers, know there will be failures and set reasonable expectations about service. Sometimes you can compensate for outages and other platform problems, sometimes you can't.
2. How would you determine how many dollars you helped the site make?