I was trying to address the issue of ROI (Return on Investment) in the project portfolio book. I don't buy project ROI. First, the idea of a project for software is an artificial construct—our consumers buy running tested features, that we happen to package in a project to release as a product. But the idea of ROI means you know how many consumers are going to buy/use the product, and you know how much you'll charge for it. That means you need a crystal ball. Mine's not working.
Instead of Producer-ROI, I've started thinking about Consumer-ROI. When someone consumes your product, what can they do with it? What kind of waste can they eliminate? What new abilities will they have?
If you think about one consumer at a time, and think about their capabilities and waste, you get a better idea of what the real ROI is. It's possible to have examples of how the product would reduce waste or increase capabilities for a given consumer. Then you ask this question: which features help solve a particular problem or reduce waste for a specific kind of consumer?
I still don't know how to predict Producer ROI (Project-based ROI), with any sort of assurance I'm telling the truth, but when I think about the users or potential users, I feel as if I'm more on the right track.