I’m struggling to provide accurate budget estimates during the initial discovery phase when requirements are still vague. What estimation techniques do you use to ensure the budget is realistic enough to get approval but flexible enough to handle scope creep later on?
3 answers
As a BA, you should never provide a single "fixed" number during discovery. I always use Three-Point Estimation (PERT). By calculating the Optimistic, Pessimistic, and Most Likely costs, you provide a weighted average that accounts for risks. This helps manage stakeholder expectations from the start. Additionally, I use Analogous Estimating by looking at historical data from similar projects we completed last year. If you document your assumptions clearly alongside these numbers, it protects you when the scope inevitably shifts during the requirements gathering phase.
Do you include a dedicated "Contingency Reserve" in your initial forecast, and if so, what percentage do you typically use for projects with high technical uncertainty?
I highly recommend "Bottom-Up Estimating." It takes longer because you have to break down the Work Breakdown Structure (WBS), but it is by far the most accurate method.
Susan is right. While it's time-consuming, the granularity of a bottom-up estimate makes it much harder for the finance committee to cut your budget.
I usually aim for a 15% contingency, but for our current AI-driven project, I’m pushing for 25%. Steven, do you find that management usually pushes back on high reserves, or do they appreciate the transparency regarding the technical risks?