Last year in the we actually asked people what was important to them. People, in particular business people, were far more interested in spending the money wisely than they were in coming under budget, for example. The point is that it's a false assumption, one foisted upon us by the traditional theory folks, that we need to accurately predict what the cost is going to be (something that we know we're not good at, and our stakeholders know we're not good at, BTW). I would rather focus on what is actually important to the business, so if they value ROI over budgeting I would choose to find ways to predict that I'll provide better ROI than other teams provide.

What business people actually want is to be able to invest their money in IT effectively, not to gamble it and hope for the best. Trying to predict up front what the budget will be actually proves to be a gamble in practice. Google the term "fixed price unethical" and you'll find an article which should get you thinking outside the traditional box.

Your point about questioning the value of investing too much in techniques which try to increase predictability is spot on. Instead of investing in bureaucracy you're much better advised to invest in techniques which increase the quality of what's being delivered (the number one issue for business people), the ROI of what you're delivering (also critical), and the chance that what you're building meets the actual needs of your stakeholders (clearly important).

Also as you imply freezing requirements is a phenomenally questionable practice. It might increase the chance of predictable costs, note the use of the word might, but it clearly increases the chance that you'll .

- Scott