Demand forecasting

Demand forecasting uses your sales history to predict how quickly each SKU will sell, so you can reorder at the right time instead of running out or over-buying.

How it works

1. Go to /app/forecasting. Retailos analyses your order history and builds a daily demand curve for every SKU in your catalogue.
2. Review the demand curves to see how sales are expected to move day by day for each product.
3. Check the predicted stockout dates shown alongside each SKU. Each prediction includes a confidence range — a best-case and worst-case window — so you know how much buffer you realistically have.
4. When a SKU is approaching its predicted stockout, Retailos surfaces a reorder recommendation. Click through to create a draft purchase order based on that recommendation. You can edit quantities before you confirm and send it to your supplier.

Good to know

• Order history is essential. The more sales data Retailos has, the more accurate the demand curves become. If a SKU has little or no order history, treat its forecast as directional rather than definitive.
• Confidence ranges matter. A wide range means the model has seen uneven or limited sales data for that SKU. A narrow range means the pattern is consistent and the prediction is more reliable.
• Draft POs save you time. The reorder recommendation pre-fills quantities in the draft purchase order, but you stay in control — review and adjust before anything is sent.
• Forecasts refresh daily. As new orders come in, the demand curves update automatically, so stockout dates and recommendations stay current without any manual input from you.

Troubleshooting

• Forecasts look flat or unhelpful. This usually means the SKU has too little order history. Keep selling and check back once more data has accumulated.
• Stockout date seems too soon or too far out. Check whether recent orders are syncing correctly. If you use Shopify, confirm your Retailos–Shopify sync is active, since Retailos reads order data from Shopify only.
• Draft PO quantities seem off. The recommendation is based on historical demand and current stock levels. If you have incoming stock not yet recorded in Retailos, adjust the draft PO quantity manually before confirming.