Inventory forecasting predicts how stock will move: when each product will run out at the expected pace of sales, and how much has to be ordered, and when, to stay in stock through the supplier's lead time. It is a sales forecast translated into units, shelf by shelf.
The forecast runs per product, because stock runs out per product. Each variant gets an expected daily pace, from its recent sales and, with a year of history, from the same weeks last year; the stock on hand divided by that pace says when it runs out, and the lead time says when the order that prevents it has to go.
Stock already on its way has to count, or the forecast orders it twice. Shopify records incoming quantities but not when they will arrive, so a forecast has to choose: count them as arriving in time, as nouz does, or as late as they plausibly can, the cautious reading for a supplier who is often late.
Saying when stock runs out is half the job. The useful forecast also prices the order, units times the unit cost, paid on the supplier's terms, and that is where inventory forecasting meets cash: the order is usually paid before a single unit of it sells.
In nouz this is the reorder planner on Insights, Inventory, for every variant whose stock Shopify tracks: what it sells a day, its days left, the day it runs out, the day to order by, the suggested order and what that order costs at today's unit cost. Variants Shopify does not track are left out and counted rather than shown as sold out, because their stock level means nothing. The Forecast page forecasts the store's orders and profit, not units per product. What nouz does not do is manage purchase orders or receive deliveries: it tells you what to order, when, and what it will cost.