Sales targets are reviewed after the books close, when nothing can be done. A run-rate projection on month-to-date sales is the usual early read, and it is fooled every month by the weekday rhythm, holiday weeks and the push of orders into the last days. A daily sales model that knows the calendar gives a probability of hitting the target from the first day of the month — and a traffic light a business team can act on.
Working with domain experts, sales were broken down by product line and the calendar effects that drive them — weekday, the end-of-month push, holiday weeks, season — were separated from the parts that cannot be forecast and should only be monitored.
Each line's daily sales are modelled as calendar effects on top of a level estimated from the last four weeks, with month-to-month persistence. Summing actuals to date and forecasts for the remaining days gives the month's total at any point, with an uncertainty that shrinks as the month fills in.
The chance of reaching the target becomes green, amber or red. Business teams see the risk while there is still a month to act, rather than a variance report after closing — the shift from after-the-fact review to proactive decisions.