This page explains the cost of forecast error in both directions: what an over-forecast, ordering more than genuine demand, actually costs compared with an under-forecast, ordering less than genuine demand, for the same style. It is for an agent or buyer coordinating a hybrid order that splits a confirmed portion from a forecast-based portion, who must decide how cautious that forecast portion should be.
At work, an agent such as Meridian Sourcing coordinating a hybrid order for Fairway Stores knows the forecast-based portion of that order could turn out either too high or too low once real demand is known. Treating both directions as equally costly is a real problem: an over-forecast and an under-forecast rarely cost the same amount for the same style, and a forecast set without weighing which direction is more expensive risks protecting against the cheaper mistake while leaving the business exposed to the costlier one. This lesson gives you the method for calculating what each direction actually costs, so a forecast can be set with the right amount of caution.
You need: lessons 3.1 to 3.3's forecasting, phasing and re-forecasting methods are useful background but not required.
By the end of this lesson you can:
Not covered here: building the forecast itself from sales history (lesson 3.1); phasing a confirmed forecast across weeks (lesson 3.2).
| Term | Meaning |
|---|---|
| Forecast risk | The risk that a sales forecast turns out to be wrong, carried by whichever company committed to that forecast. |
| Over-forecast | A forecast that turns out higher than genuine demand, resulting in surplus stock once the real selling period ends. |
| Under-forecast | A forecast that turns out lower than genuine demand, resulting in missed sales once the real selling period shows stronger demand than planned. |
| Asymmetric cost | A situation where the cost of one direction of forecast error is genuinely different from the cost of the other direction, for the same style. |
| Hybrid production | A factory building part of a run to its own forecast before a buyer confirms it, and finishing the rest once the buyer's own order is confirmed. |
| Markdown | A reduction in the shop price of goods that have not sold as forecast, used to clear the remaining stock. |
Every forecast carries forecast risk in two possible directions: an over-forecast, producing more than genuine demand, or an under-forecast, producing less. These two directions are not automatically equal in cost. An over-forecast's surplus is normally resolved through a markdown, losing only the gap between full price and markdown price on the surplus pieces. An under-forecast's missed pieces are normally lost sales entirely, losing the full margin on every piece that could have sold but was never produced. Because these are different calculations, one direction is often genuinely more expensive than the other for the same style, an asymmetric cost that should shape how cautious the forecast is set.
| What can differ | The shapes you will meet | Why it matters |
|---|---|---|
| Export, hybrid, agent-coordinated | Meridian Sourcing weighs the cost of Fairway Stores' 5,000-piece forecast portion turning out too high against it turning out too low, before agreeing how much of it Delta Garments should build ahead of confirmation. | An agent coordinating a hybrid split must set the forecast portion's own size with both cost directions in view, not just the more familiar over-forecast risk. |
| Domestic, MTS, continuous | Cedar & Co. weighs an over-forecast's markdown cost against an under-forecast's lost sales continuously, since its standing forecast is checked and adjusted every period. | A Make to Stock buyer meets this same cost comparison repeatedly through a season, not once before a single order. |
| A style with a high markdown rate | A fast-fashion style with a steep markdown rate makes over-forecasting relatively more costly compared with a basic style carried at a shallower markdown. | The specific markdown rate a style carries changes which direction is more expensive; it is not the same for every style. |
| A style with a short selling window | A style with a short sell-through window, replaced quickly by the next style, makes under-forecasting more costly, since a missed sale in that narrow window may not be recoverable later in the season. | A short selling window raises the real cost of an under-forecast, since there is little time left in the season to make up a missed sale. |
Fairway Stores' order for Order A's own T-shirt is structured as a hybrid: 10,000 pieces confirmed outright, and a further 5,000 pieces built by Delta Garments against a forecast before Fairway Stores fully confirms them. Meridian Sourcing calculates what happens if that 5,000-piece forecast portion turns out 800 pieces too high, against what happens if it turns out 800 pieces too low, using Order A's own fixed figures.
| Question | 800 pieces too high (over-forecast) | 800 pieces too low (under-forecast) |
|---|---|---|
| What happens to the 800 pieces? | Produced, delivered, but unsold at full price; cleared at Order A's own 20% markdown | Never produced under this order; the demand for them exists but is not supplied |
| What is actually lost? | The gap between the shop price and the markdown price on 800 pieces | The full margin Fairway Stores would have earned on 800 pieces that could have sold |
| Which figure is larger? | Confirmed in the worked number below | Confirmed in the worked number below |
Meridian Sourcing calculates each direction separately using Order A's own fixed figures: shop price USD 19.90, wholesale USD 8.00, and a 20% markdown of USD 3.98 per piece.
A common mistake is assuming that being wrong by the same number of pieces in either direction costs the same amount, and so aiming for the forecast that feels safely in the middle without calculating either direction's real cost. If Meridian Sourcing had set Fairway Stores' 5,000-piece forecast portion purely to minimize the chance of any surplus, without calculating that an under-forecast actually costs nearly three times as much per piece as an over-forecast for this style, the forecast would be protecting against the cheaper mistake while leaving the business more exposed to the costlier one.
When a forecast turns out wrong, it can be wrong in two different directions: too high, leaving surplus stock, or too low, missing sales that existed. These two directions usually do not cost the same amount for the same style. Too high normally means marking down the surplus, losing only the gap between the full price and the markdown price. Too low normally means losing the entire margin on every piece that could have sold but was never made. For Order A's own figures, an under-forecast costs nearly three times as much per piece as an over-forecast of the same size. A forecast should be set with that real cost difference in mind, not aimed evenly at avoiding both directions as if they cost the same.
Fairway Stores' buyer, wary of an over-forecast leading to a markdown, asks Meridian Sourcing whether the forecast portion should be reduced from 5,000 to 4,200 pieces.
Should Meridian Sourcing recommend reducing the forecast portion from 5,000 to 4,200 pieces to reduce markdown risk? Decide before you open the answer.
No. Reducing the forecast portion protects against the cheaper of the two risks, an over-forecast at USD 3,184 for 800 pieces, while increasing exposure to the costlier one, an under-forecast at USD 9,520 for the same 800 pieces.
First, why this specific style makes under-forecasting the costlier direction. Order A's own fixed figures give a markdown gap of USD 3.98 per piece against a full margin of USD 11.90 per piece, nearly three times as much; this is a property of the style's own numbers, not a general rule for every style.
Second, what the buyer's request actually trades. Lowering the forecast portion by 800 pieces removes at most USD 3,184 of possible markdown cost but adds at most USD 9,520 of possible lost-margin cost if demand reaches the original level, a trade that moves the expected cost in the wrong direction.
Meridian Sourcing should hold the forecast portion at 5,000 pieces, since this style's own numbers show under-forecasting as the more expensive direction, and recommend that any caution be applied through phasing and re-forecasting during the season, covered in lessons 3.2 and 3.3, rather than through cutting the forecast portion itself.
Framed as the two outcomes Fairway Stores' buyer is actually choosing between, using the figures already calculated in this lesson.
Cutting the forecast portion by 800 pieces does not remove risk; it exchanges a smaller, well-understood markdown risk for a larger, less visible lost-margin risk, a net worsening of USD 6,336 in the worst case for this specific style's own figures. The saving a nervous buyer sees in avoiding a markdown is smaller than the lost-margin exposure the same decision creates.
The workbook's "Forecast cost comparison" sheet takes a style's own shop price, wholesale price and markdown rate, calculates the cost of an over-forecast and an under-forecast for any given piece count, and flags which direction is more expensive for that specific style.
Meridian Sourcing's own row records Order A's own figures against an 800-piece error in each direction, with the sheet calculating USD 3,184 for over-forecast, USD 9,520 for under-forecast, and flagging under-forecast as the costlier direction.
1. Why do an over-forecast and an under-forecast for the same style usually cost different amounts?
2. Using Order A's own figures, USD 3.98 markdown and USD 11.90 margin per piece, what is the cost of an 800-piece error in each direction?
3. Fairway Stores' buyer wants to cut the forecast portion to reduce markdown risk, even though under-forecasting costs nearly three times as much per piece for this style. What should Meridian Sourcing recommend?
| EN | AR | FR | TR | Meaning |
|---|---|---|---|---|
| Forecast risk | مخاطر التوقع | Risque de prévision | Tahmin riski | The risk that a sales forecast turns out to be wrong, carried by whichever company committed to that forecast. |
| Over-forecast | فرط التوقع | Surestimation de la prévision | Aşırı tahmin | A forecast that turns out higher than genuine demand, resulting in surplus stock once the real selling period ends. |
| Under-forecast | نقص التوقع | Sous-estimation de la prévision | Eksik tahmin | A forecast that turns out lower than genuine demand, resulting in missed sales once the real selling period shows stronger demand than planned. |
| Asymmetric cost | تكلفة غير متماثلة | Coût asymétrique | Asimetrik maliyet | A situation where the cost of one direction of forecast error is genuinely different from the cost of the other direction, for the same style. |
| Hybrid production | الإنتاج الهجين | Production hybride | Hibrit üretim | A factory building part of a run to its own forecast before a buyer confirms it, and finishing the rest once the buyer's own order is confirmed. |
| Markdown | تخفيض السعر | Démarque | İndirim (markdown) | A reduction in the shop price of goods that have not sold as forecast, used to clear the remaining stock. |
You now know the real cost of a forecast that is wrong in either direction. The final lesson brings forecasting, phasing, re-forecasting and cost weighing together into one complete worksheet. Back to the course map