In my post Committing to Estimates, I presented Cone of Uncertainty with my thoughts and stated we need to be pragmatic about the depth to which we should go into requirements analysis before we commit to estimates. I received offline feedback from quite a few people that going down to the right depth consumes 20% of the project cost - something which business is unwilling to spend just to get scope defined and mitigate risks of cost escalations.
Businesses need to realize spending 20% of project costs to get scope defined unambiguously is a better deal than seeing a failed project due to premature commitments. There is a chance after investing to get the scope sorted out you realize you cannot afford the cost of the project, but you at least have an option to rethink your strategy, work out a solution with your IT team or IT partner that can be delivered within your budget... you have the opportunity to apply the Iron Triangle!
Going Agile resolves the concern of investing in scoping and focuses on delivering the product from very early stages of investments, however it doesn't resolve the concern of getting an upfront estimate for budgeting IT costs. It is important that we as business analysts educate the business that there are no silver bullets in IT projects and work with them to determine a right equilibrium.
At the same time, it is essential that BA community comes up with innovations to reach unambiguous scope in lesser time and effort. Tools that help focus on analysis and automatically generate formatted documentation, tools that help bridge onsite/offshore divide with powerful collaborative features, tools that are easy for business users to grasp and help them do effective reviews and commenting within the tool, wireframing tools that provide visualization of end-to-end functional flow......
I am trying to hunt for such tools and will keep posting my assessment whenever I find a good one, and I welcome your feedback about tools that have worked best for you.