A sales forecast is an estimate of what a store will sell in the weeks and months ahead, in orders and revenue. The simplest honest method starts from the store's own past: the same weekday last year, scaled by how the recent weeks compare with the same weeks a year earlier.
Forecasts come in two families. Statistical models fit a curve to the history and extend it; they can be precise, and they are hard to check, because nobody can say which past week drove which future one. The other family reuses the store's own calendar: every future day is the same weekday a year before, so a Saturday stays a Saturday, a sale season lands where it fell last year, and every figure can be traced back to orders that really happened. Forecasting sales from your own orders walks through that method step by step.
The past then needs one correction, for how the store has moved since. Comparing the last few months of orders with the same months a year earlier gives a growth rate, and a careful forecast caps it so one unusual month cannot run away with the whole year. With less than a year of history there is no season to copy, and the honest fallback is the recent pace repeated for a few months, said out loud.
A sales forecast says how busy the store will be. It cannot say whether those months are worth having, because the same revenue can arrive with very different costs; that question needs a profit forecast, the same months priced through every cost.
That is how the Forecast page in nouz works: the last 364 days of orders move forward to the same weekday, scaled by up to the last 91 days against the same days a year earlier, with the rate held between a quarter and three times. With fewer than 28 days of orders there is no forecast at all; with less than a year, the last four weeks repeat for this month and the next three, and the page says so. The help centre sets out how the forecast is built, rule by rule.