If, say, further insulating your house or building a sand battery will pay for itself in 50 years, it's a bad investment, financially speaking, and probably environmentally speaking as well. You can deploy "the same amount" of resources in something else with a higher ROI, like maybe solar panels with a one-year payback, and get a much bigger benefit. This is an important consideration as long as you are constrained by some kind of resource limitation.
So I think ROI is a first-order consideration.