How Inventory Demand Planning Actually Works

What Inventory Demand Planning Actually Means

The phrase covers two jobs that get mashed together constantly. Demand planning is the forecast: how many units of a SKU will sell, and in what window. Inventory planning takes that forecast and turns it into a decision — how many units to hold, and when to trigger the next order. A third term, inventory management (sometimes called inventory control), is neither of those — it's tracking what you physically have, where it sits, and in what condition.

These three feed each other in one direction: the demand forecast sets the inventory plan, and the inventory plan sets what the day-to-day management system is supposed to be tracking against. When people ask about "demand planning and inventory optimization" or "inventory and demand planning" as if they're separate disciplines, they're usually really asking one question: how much should I be holding, and when do I reorder it. That's the reorder point, and it's a formula, not a guess.

The Formula, With a Worked Example

The core equation behind almost every inventory planning decision is this:

Reorder point = (average daily sales × lead time in days) + safety stock

Say a SKU sells 20 units a day on average, and the supplier's lead time — order placed to stock landing — is 45 days. You've decided to hold 15 days of extra demand as a buffer against variability. The math: (20 × 45) + (20 × 15) = 900 + 300 = 1,200 units. That's the level at which the next purchase order has to already be placed, not merely being discussed, or you run out before the replenishment arrives.

Now change one input. A promotion or a seasonal spike pushes daily sales to 30 units — a 50% jump that's common and easy to miss in a monthly review cycle. The same 1,200-unit buffer now covers roughly 40 days instead of 60. Nothing in the ledger looks wrong yet, but the safety margin is eroding fast, and by the time a manual review catches the change, the reorder may already be late. This is the entire argument for reviewing triggers more often than you review the plan.

When the Forecast Turns Out Wrong

A demand plan being wrong isn't a failure of the model — it's the normal case. What matters is what happens next.

• Selling faster than forecast: confirm it's a real shift (a promotion, an ad push, a viral moment) and not a blip, then choose between expediting the existing PO at extra freight cost or accepting a short gap and protecting margin. Panic-ordering at any price is rarely the right call.
• Selling slower than forecast: the cash is now sitting in stock that isn't moving. The decision becomes whether to hold it, discount it, or make the removal-versus-liquidation call before storage fees compound the loss.
• The forecast says one thing, the physical count says another: stop and check the input before touching the model. One mistake we've made ourselves: trusting the inventory ledger over a physical check. The system showed adequate stock and no reorder due for three weeks — but the units were sitting in an unprocessed inbound shipment, received by the fulfillment center but not yet checked into sellable inventory. The forecast wasn't wrong. The number it was fed was.

Where This Gets Amazon-Specific

Generic demand planning treats a stockout as a lost-sales problem. On Amazon it's also a rank problem — a SKU that goes out of stock mid-campaign loses organic position and buy box share that doesn't come back the moment stock lands again. Inventory planning here has to account for restock limits, IPI-linked storage constraints, and FBA lead times that move around more than a standard supplier lead time, because inbound processing speed at the fulfillment center is its own variable.

Across more than $500M in managed revenue across 100+ brands, the same pattern shows up repeatedly: the forecast was fine, but the reorder trigger didn't account for one of these Amazon-specific frictions — a longer-than-usual check-in time, a storage limit that blocked a full replenishment, or aged-inventory surcharges eating the margin on stock that was technically "in the plan." A cookware brand carrying seasonal spikes around holiday gifting, for instance, needs safety stock sized for that window specifically, not an annual average that smooths the peak away.

Where Dr. Stock Fits

Dr. Stock, a managed product from Fable 5 at Full Circle, works the inventory side of this: the reorder timing behind stockouts, cash trapped in slow SKUs, storage and aged-inventory fees, FBA fee errors, and the shipment discrepancies that make a ledger disagree with the warehouse. Inventory purchasing decisions always go to a human, regardless of how much autonomy the client sets elsewhere, and Orbit's trackers come with it at no extra cost. If the actual leak is on the ad side — spend driving demand faster than the forecast expected — that's a Dr. PPC problem, not an inventory one. Either way, the formula above is the same one any tool you use should be running.