No matter how good you think you are, you're not that good. No one is; your crystal ball is just as cloudy as anyone else's. This is the single biggest reason IT projects have such a high failure rate. Remember: The cost of IT initiatives will typically exceed original estimates by an average of 100 percent.
Institutional knowledge is lacking as to the result of major initiatives, the advice and counsel of IT is routinely omitted or ignored, and business process change relies too heavily on IT ownership of those business processes. Frequently an estimate is called for, if not to virtually guarantee, a project cost before the scope has been fully defined.

As an IT professional, whatever your role on a project, you must provide business managers with parameters for setting funding expectations and force those business managers to explain why their assumptions are valid.

If you're an IT manager, track all major development efforts throughout the enterprise and regardless of your role, participate in the creation of a knowledge base of maintenance and support costs to drive future verifiable and credible estimation. Don't underestimate the future costs of maintenance and support and whatever you do, don't make the classic cardinal error: Do not, under any circumstances, pad budgets in anticipation of an underestimation. Keep track of project costs as the project unfolds and communicate, immediately and vociferously, the instant you detect even the potential for an overrun.