My cofounder from 2010-12 is an amazing product manager. Here's he giving a 5-minute talk on “How $40 Saved Us 9 Months and $2MM” https://vimeo.com/24749599
The short version is that we were looking at building a complex social experience that had Facebook newsfeed items as a key component. We used GreaseMonkey to fake it entirely: user testers would log in to Facebook on our tricked-out computer, a GreaseMonkey script would steal faces and names from below the fold, and then insert fake news items into their feed that were apparently from their real friends. (At the end of the user test we'd come clean so they weren't confused.)
What we learned is that people hated the idea. So that meant I never had to write a line of production-grade code for this. Some of our competitors had very similar ideas and ended up building full products to learn the same lesson. What they built was way more scalable, of course. But that's not a virtue in a code base that just gets thrown out.
The short version is that we were looking at building a complex social experience that had Facebook newsfeed items as a key component. We used GreaseMonkey to fake it entirely: user testers would log in to Facebook on our tricked-out computer, a GreaseMonkey script would steal faces and names from below the fold, and then insert fake news items into their feed that were apparently from their real friends. (At the end of the user test we'd come clean so they weren't confused.)
What we learned is that people hated the idea. So that meant I never had to write a line of production-grade code for this. Some of our competitors had very similar ideas and ended up building full products to learn the same lesson. What they built was way more scalable, of course. But that's not a virtue in a code base that just gets thrown out.