After 25 years of helping organizations manage their project portfolios, I think it's safe to say that the single biggest stumbling blocks most organizations have in implementing Agile Project Portfolio Management is canceling projects.
Early on in my career, my first assignment at a large telecom company was to make a recommendation as to whether or not the company's biggest R&D project should be cancelled. Knowing full well that whatever I reported back could end up costing me my job, I spent a couple of weeks talking to everyone associated with the project, from senior management to rank-and-file engineers. It became clear that the project was an innovative idea that didn't work and was never going to produce an economical product. The company had learned a lot during the project that could be reused, but the technology they had invested in just wasn't the right approach.
How a company handles this realization tends to vary considerably by industry. My clients in the pharmaceutical industry cancel projects like this as a normal cost of doing business. They know that not every research project they initiate will end up producing a marketable drug. With high-tech electronics and software companies, the acceptance of this level of inevitability doesn't come quite so easily.
For example, when I was interviewing the technical staff at the telecom company, one