72. How would you assess whether a new forecasting method is actually better than the current one?
HardSales process
Technical
Show model answer
Compare methods using the same historical cutoff, horizon, eligible opportunities, and outcome definitions, preserving what was known at the time. A reconstruction that uses later buyer information gives the new method an unfair advantage.
Examine bias, error, concentration, and performance across relevant cohorts, not just a single good quarter. Use a named baseline and an appropriate prospective check before claiming the method improves planning or replacing the existing workflow.
What interviewers listen for
- Prevent later information from leaking into comparisons
- Compare equivalent horizons and definitions
- Assess bias and cohort performance against a baseline