I've been thinking about how to manage the project portfolio, and I just realized why so many project portfolio efforts fail.
There are three kinds of cycles the project portfolio managers need to manage:
- Project cycles: when the project could release something
- Planning cycles: how often the management team assesses the project portfolio
- Business cycles: when customers want something new
To actually manage the portfolio, the project cycles have to be less than (or equal to) the planning cycles. The planning cycles have to be less than (or equal to) the business cycles (unless you don't care if you only react to customer requests instead of planning for them). Sometimes, you do want to react to customer requests (have a customer request trigger a planning cycle), but once you have a relatively mature product, the customer requests don't always align with your product roadmap.
To be most flexible, the Agile lifecycles shine for managing the project portfolio. If you can't manage an Agile lifecycle, use an incremental lifecycle. You'll have more flexibility on when to end the project–which means you can actually manage the project portfolio.