A yearly average hides the months that matter. If you order to the average, you run out in peak season and sit on stock in the slow months. This guide shows how to measure seasonality from your own sales, turn it into a forecast, and time your orders around it.
Find your seasonal pattern
Start with at least a full year of monthly unit sales per product, and two or three years if you have them. Work out the average month, then compare each month to it. The result is a seasonal index.
- Average month: Total units for the year ÷ 12
- Seasonal index: Units sold in a month ÷ the average month
- Above 1.0: A stronger month than usual
- Below 1.0: A weaker month than usual
Work it through
Say a product sold 1,200 units last year, so the average month is 100. November sold 220 and February sold 60.
| Month | Seasonal index |
|---|---|
| November: 220 ÷ 100 | 2.2 |
| February: 60 ÷ 100 | 0.6 |
To forecast next November, take your current baseline and multiply by the index. If the product now averages 130 units a month, November's forecast is 130 × 2.2, or 286 units.
Separate season from noise
One year of data can mislead. A stockout, a sale or a single viral post will show up as a seasonal dip or spike that will not repeat. Clean the history before you trust it.
- Stockouts: A month with no stock shows low sales, not low demand
- Promotions: Note which months had discounts or heavy ad spend
- One-off events: Leave them out of the index
- Trend: If the product is growing, compare each month to the average for its own year
Back-plan from the peak
The order for peak season is placed well before the peak. Count backwards from the date stock needs to be on the shelf, using your supplier's lead time and any extra time for shipping and receiving.
- In-stock date: When peak demand begins
- Minus lead time: Production plus transit
- Minus receiving time: Days to check in and shelve stock
- Result: The latest date you can place the order
Lead times often stretch before busy periods, so ask your supplier for their cut-off dates early.
Adjust your reorder point by season
A reorder point built on average daily sales is too low in peak season and too high after it. Recalculate it with the forecast for the period ahead, not the period behind. The same applies to your buffer.
- Before peak: Raise the reorder point and safety stock
- During peak: Review weekly, and compare actual sales to forecast
- After peak: Lower both, so you do not reorder into a slow season
See how to calculate a reorder point and how to calculate safety stock for the formulas.
Doing this at scale
StockSense forecasts per-SKU demand from your order history using velocity, trend, seasonality and lead-time-aware safety stock. It recommends a reorder point, quantity and date for each SKU, ranks SKUs by revenue at risk, and drafts purchase orders grouped by supplier. See StockSense.