I read Sayonara Sony: How Industrial, MBA-Style Leadership Killed a Once Great Company, and was struck by how the company appeared to make decisions—what will make us the most money now?
If you make your project portfolio decisions based on money, estimated cost, you too will have Sony's problems. This is why cost is the wrong question to ask. You want to ask the value question, “How much value will this project provide?”
This is not a return on investment question. That's a prediction question. You can't answer that question either.
When you use cost as the “which project should we fund?” you can guess right. But you guess based on what you think the market will buy, not what you decide is the right strategy for your company. Big difference.
This is why you want a mix of projects in your project portfolio: Projects that keep the lights on; projects that keep cash coming in; and projects that are innovative, that have the power to transform your organization. And the problem is this: in software, you often can't tell which one is which. This is why agile or lean is so wonderful. You don't have to be able to tell.
If you use incremental budgeting and an agile approach to your project portfolio, and an agile/lean approach to your projects, you can pivot when you need to.
Now, you might need to do a quick SWAG (Scientific Wild Tush Guess) for an order of magnitude estimate for your project. Is this project bigger than a breadbox? Is it smaller than a space station? That's not a project estimate. That's a SWAG, an order of magnitude estimate. You do a SWAG with a few people in under an hour. Two hours max.
For the project portfolio, you make the evaluation decision based on your best guess of the project's comparative value to the organization. This is why it's so helpful to be able to deliver something in a short period of time, such as 2-6 weeks, when the project portfolio people will want to reevaluate the portfolio again.
Oh, and those Sony engineers? The ones who were laid off when Sony couldn't keep their commitment to lifetime employment? Some of them went to LG. You know, the company that you probably bought your flat screen TV from? All I know, is that every hotel I stay in has an LG TV. In our house, we have (smaller) LG TVs.
When you use value to make project portfolio decisions, you decide on the most valuable project. If you change your mind, you don't can the people. You change what the people do—you flow work through the team. Is this more difficult with hardware? Of course. Is it impossible? No.
People want autonomy, mastery, purpose. (Yes, this is from Dan Pink.) Give it to them. Tell them what is the number one project. The number two project, etc. Tell them what is not funded, and why. They will work for you, to make it happen.