Inventory Demand Forecaster
Inventory Demand Forecaster is a reusable ecommerce AI prompt for Shopify, Amazon, WooCommerce sellers. Use it to turn cleaned sales and supply inputs into a…
What this prompt helps you do
Turn sales history into inventory forecasts with reorder recommendations. Accounts for seasonality, promotions, and lead times.
Prompt text
Act as an inventory planning analyst for an ecommerce business. Given sales history data, generate: 1. DEMAND FORECAST: - 30-day projection by SKU - 90-day projection for top sellers - Seasonality adjustments (month-over-month patterns) - Confidence intervals (high/low estimates) 2. REORDER RECOMMENDATIONS: - Which SKUs to reorder this week (stockout risk < 30 days) - Optimal order quantities (EOQ calculation) - Lead time buffer recommendations - Minimum order quantity considerations 3. RISK ASSESSMENT: - Products at risk of stockout (ranked by revenue impact) - Overstock items (> 90 days supply) - Dead stock identification (> 180 days) 4. PROMOTION IMPACT: - Expected demand spike for planned promotions - Safety stock adjustment for promo periods - Post-promotion demand dip estimate 5. CASH FLOW IMPACT: - Estimated inventory investment needed this month - Working capital tied up in slow-moving stock - Liquidation recommendations for dead stock Focus on actionable reorder dates and quantities. Flag anything that requires immediate attention with ⚠️.
Turn cleaned sales and supply inputs into a reviewable reorder proposal rather than an automatic purchase order.
Before you start
- Export at least 12 months of SKU-level sales when available.
- Add current stock, open purchase orders, supplier lead time, MOQ, returns, and planned promotions.
- Agree on the service-level target, stockout cost, and cash limit.
How to test it safely
- Backtest the last eight to twelve weeks.
- Compare with a simple recent-sales baseline.
- Track forecast error, stockouts, excess days of supply, and cash tied up.
Operator-ready setup plan
Clean and label the data
Use consistent SKU, date, unit, and currency fields. Separate returns, stockouts, bundles, and one-off promotions.
Checkpoint: Missing periods and abnormal events are visible rather than silently treated as normal demand.
State the operating assumptions
Provide lead time, order cadence, MOQ, safety-stock policy, promotion calendar, and known supply constraints.
Checkpoint: The recommendation can be traced to explicit assumptions.
Request a range
Ask for base, low, and high demand cases with the formula or reasoning behind reorder dates and quantities.
Checkpoint: The output shows uncertainty instead of one false-precision number.
Backtest before buying
Run the same method on an earlier period and compare the forecast with what actually sold.
Checkpoint: Forecast error is understood by SKU group.
Approve orders outside the model
Review cash, supplier, storage, margin, and promotion constraints before creating a purchase order.
Checkpoint: A named inventory owner signs off on every order.
Limits to account for
- The Prompt is not covered by the current deep-source snapshot.
- Language models can make arithmetic or table-handling errors.
- Sparse history, stockouts, new SKUs, and promotions can make a forecast unreliable.
Questions at this experience level
How much sales history do I need?
Twelve months is preferable for seasonality, but the useful amount depends on product life, promotions, and data quality. Label gaps and special events.
Can this create purchase orders automatically?
It should first produce a proposal. Keep supplier commitment and cash allocation behind human approval.
How do I know whether the forecast is useful?
Backtest it against actual sales and compare it with a simple baseline. Track errors by SKU group, not only one store-wide average.
Author: Author information not found
Content checked: