Diminishing returns is one of the most misunderstood concepts in engineering.

Most devs know they can do 80% of the work very fast, but that last 20% will consume their soul and 10 years of their life. Because of that, their minds think it's not worth investing in the project anymore: "I could be more productive in this new side project". Which is true, but that also means that you have two half-baked projects that nobody can use because they are not finished. You fail on both at the same time.

In the market, diminishing returns are diminishing returns for everybody. To have the best product, you focus on diminishing returns exactly because your competitor will phase out and give up: "it is too much money for little gain". Any rational player will decide against it. But, if you truly want to win, you must keep pushing your product's edge up. You must work through diminishing returns profitably. If you don't, you are already gone, and your competitors are going to love it.

Learning the law of diminishing returns is not done so that you can avoid it when it happens. It's done so that you can go through it in a way that makes sense.