The existing product was a C code base that had been hacked upon for 16 years (!), with several distinct breaks in continuity. Sufficiently bad that it took hours of investigation to make the slightest change safely, and its performance was increasingly unacceptable, plus it was buggy. I've done a
lot of "code archaeology" before and some since, and this is the first time I decided a code base was a total loss with no insurance. So the deadly total rewrite was planned.
The business model was "gated open source", as in paying customers got source, the users were almost without exception system administrators, not programmers, and with one exception who was something of a programmer (at least he did more with the awful code base than any other user) and willing to give it a try, they demanded the "extension" language (actually the complicated business logic) be in Perl ... despite to our knowledge at there were no successful examples of this, whereas Lisp has many, with EMACS and AutoCAD being famous wild successes. (The whole thing couldn't be in Perl because of performance requirements.)
Despite being the only programmer working on it and not seriously experienced with Perl, I got overruled. The CEO, the only manager, was the sort who during this time of decreasing revenue felt compelled to move from Class B to Class A office space (https://en.wikipedia.org/wiki/Office#Grading), and more of it---yet somehow the 3 technical people supporting almost all revenue got allocated only one office, which was tight for 2 people, and I'm one of those who needs quiet for maximum productivity (the other two were sysadmins and technical support, and had to do a lot of the latter over the phone. Whine, whine.)
Basically, only with maximum productivity on my part was this possible to pull off (and I've done this sort of thing before), as the users were years overdue for real technical fixes and progress (long before I arrived).
The board got really upset, the executive director blamed me and they believed her (I after all had been there for less than 2 years and had been able to do only one month of programming in my first year due to Y2K and other sysadmin and tech support demands) ... I had to bail.
It was partly my fault, I was not very productive for several months after my major architectural decision was vetoed and that work was dumped, which happened roughly the time the office move made the company's financials likely terminal, and I should have outright refused the CEO when she demanded I drop everything to help try to sell more of the old (which her fantasy budget demanded) and pre-sell versions of the new (then again I have to wonder if that would have made any difference).
The company died an ugly death, but not before firing in revenge a close friend I'd hired to do system administration and tech support.
Bleah. Bottom lines: avoid recruiting your friends because the company may later go to hell, if you observe the former then don't stay at places that don't respect programmers, can't keep them, and who don't have their eye on the ball of what brings in the cash. None of this was apparent when I accepted their offer, but I'm sure most of us know that story.