Pareto's rule, also known as the 80-20 rule, tells us that we can acheive 80% of the benefits from 20% of the software. Ron Jeffries examines the implications of this rule, showing that the return on investment (ROI) of building 80% of five systems is higher than the ROI of building 100% of a single system.
Realistically, it might not be possible to always start at the top 20%, but you very likely could stop with the top 30-40% if you were flexible. As we see in Examining BRUF, traditional teams would apparently be better off if they were to only attempt the top 55% of the functionality of a system as the other 45% proves to be a complete waste. The point is that Ron is presenting an argument that we should all take seriously.