I use 80/20 rule quite a bit of coding. Developers can easily get obsessed with minutia that doesn't help the product organization or has ample diminishing returns. Tuning a business app is a great example. Is 80% performance tuning good enough? Most times, yes. In fact 80% is probably too much in most cases. I can spend the remaining 80% working on new features that would greatly enhance the business' service.
Only in rare exceptions would I want the inverse: spend the remaining 80% of my effort to gain the final 20% improvement.
Only in rare exceptions would I want the inverse: spend the remaining 80% of my effort to gain the final 20% improvement.