Demand forecasting
Reduce time spent on demand planning. Pick the horizon that fits your operation, see confidence bands, drill into any SKU.
- What it uses
- Sales history per SKU (item ID, date, quantity, price).
- Minimum one year of history is preferred for accuracy.
- Upload as CSV, or let MLmargin pull data via Shopify or your existing ERP.
- What you get
- Daily, weekly, biweekly, or monthly forecasts.
- Drill-down on each SKU for a focused view.
- Seasonality recognition for fashion, holiday, and weather-driven categories.
- Trend interpolation that adapts to your growth or decline.
- How it works
Deep learning where the data is dense, classical ML where it's sparse — combined within a Reinforcement and Transfer Learning loop. The model adapts to your data, not the other way around.