Your December storage bill is calculated from your November inventory. By the time you see the number, the decision that caused it is six weeks old.
Most Q4 inventory advice is about not running out. That is half the problem and the easier half. The other half is that Amazon has spent several fee cycles building a structure specifically designed to make warehousing behaviour expensive, and Q4 is when all of it bites hardest at once.
Three separate charges can apply to the same cubic foot in the same month. The base rate roughly triples. The assessment lags by a month, so the feedback loop is too slow to correct inside the quarter. And the thresholds that trigger the worst of it are calculated on trailing thirteen-week averages, which means the damage is done before the metric moves.
This is the arithmetic, the forecast build, and the deadline that actually matters.
Fee stacking — the simultaneous application of multiple independent storage charges to the same inventory. On Amazon a single slow-moving unit in December can incur the peak base storage rate, an aged inventory surcharge and a storage utilization surcharge concurrently. None of these replaces or offsets another, which is why holding cost per cubic foot can be several times the headline rate.
Why Inventory Math Is Different in Q4
- The base rate roughly triples from October through December.
- Charges compound rather than substitute. Peak rate, aged surcharge and utilization surcharge apply independently.
- Assessment lags a month. Fees are charged on the 15th based on the previous month's daily average cubic footage, so you are always paying for a decision already made.
- Thresholds use trailing averages. The utilization ratio is computed over thirteen weeks, so it moves slowly in both directions — including when you try to fix it.
- Peak fulfillment fees run to mid-January, so January clearance sells at reduced prices while still paying elevated per-unit costs.
- Capacity limits constrain the correction. Selling out is not automatically fixable if your restock capacity is capped.
The asymmetry that should drive your decisions
Running out of a good seller in November costs you the margin on units you could have sold. Over-ordering costs you triple storage, potentially two surcharges, tied-up capital through the most cash-hungry quarter, and a January carrying a drag into the new year.
Those are not symmetric, and the direction of the asymmetry depends entirely on your margin. High-margin products should err toward over-stocking; thin-margin bulky products should err toward running lean, because the storage cost per unit of margin protected is completely different.
The Base Storage Rate Card
| Size tier | Jan–Sep | Oct–Dec | Multiple |
|---|---|---|---|
| Standard-size | $0.78 / cu ft | $2.40 / cu ft | 3.1x |
| Oversize | $0.56 / cu ft | $1.40 / cu ft | 2.5x |
| Dangerous goods, standard | $0.99 / cu ft | $3.63 / cu ft | 3.7x |
| Dangerous goods, oversize | $0.78 / cu ft | $2.43 / cu ft | 3.1x |
How the charge is actually computed
Not on unit count. On the daily average cubic footage your inventory occupied across the prior month, assessed on the 15th. Two implications follow that most sellers miss.
First, packaging dimensions are a storage cost lever. Reducing a carton by half an inch across a few thousand units changes your cubic footage and therefore your bill, every month, permanently. That is worth more attention than most sellers give it and it compounds with the size-tier effect on fulfillment fees.
Second, timing within the month matters. Inventory that arrives on October 2 pays essentially a full month of peak-rate storage. The same inventory arriving October 25 pays a fraction of it. When you are close to a deadline anyway, arriving later within the allowed window is genuinely cheaper — provided you do not miss the cutoff, which costs far more than the storage saved.
Pull your actual daily average cubic footage from the Storage Fees report rather than estimating from unit counts. Brands are routinely surprised, usually because one bulky slow mover occupies a disproportionate share of the total while representing a small share of revenue.
The Three Surcharges That Stack
$2.40 per cubic foot standard-size, October through December. Applies to all inventory regardless of age or velocity.
Applies once a unit passes the age threshold, escalating in tiers the longer it sits. Charged per cubic foot or per unit, whichever is greater.
Applies where daily volume is at least 25 cubic feet and the utilization ratio exceeds 22 weeks. Only on inventory aged over 30 days.
A bulky, slow-moving, over-stocked SKU in December is the worst case in the entire fee structure. It is also the easiest to identify in advance.
Amazon has revised these more than once and published sources currently disagree about whether the first surcharge tier begins at 181 or 271 days, and about the exact rates at the upper tiers. What is not in dispute is the direction: the thresholds have tightened, the upper tiers have become more expensive, and a new tier was added at the top for 2026. Pull your own numbers from the Inventory Age report and the Storage Fees report in Seller Central rather than trusting any published table, including this one.
Who is exempt from the utilization surcharge
- Sellers whose average daily inventory volume is below 25 cubic feet.
- New sellers, defined as those whose first shipment to a fulfillment center was within the past 365 days.
- Individual account sellers.
- Dangerous goods inventory.
If you are close to the 25 cubic foot line, that threshold is worth managing deliberately rather than crossing by accident.
The Utilization Ratio, Calculated
The metric that decides whether the third surcharge applies to you, and one most sellers have never computed.
Reading the number
The ratio is essentially weeks of cover measured by volume rather than by units, which matters because it weights your bulky SKUs far more heavily than a unit-based weeks-of-supply figure does. A catalog that looks healthy on units can breach on volume because of one large slow mover.
The practical management rule
Start slowing replenishment at around 18 weeks, not at 22. Because the ratio is a trailing thirteen-week average, it responds slowly, and by the time you read 22 you have already been accruing the condition for weeks. Treating 18 as the intervention point gives the average time to move before the threshold is crossed.
Track it per SKU where you can, not just at account level. Account-level health routinely conceals one or two ASINs quietly driving the whole charge.
IPI and Capacity Limits
Your Inventory Performance Index is scored 0 to 1,000 on a rolling basis and determines how much you are allowed to send.
The four inputs
- Excess inventory percentage — how much of your stock exceeds a healthy cover level.
- Sell-through rate — units sold against average inventory held.
- Stranded inventory — units in the network with no active listing. The cheapest of the four to fix and the most frequently ignored.
- In-stock rate — availability of your replenishable ASINs.
Why it matters more before Q4 than during it
A low IPI triggers storage capacity restrictions, and restrictions bite hardest when you need to send peak inventory. The score moves slowly, which means September is roughly the last month where improving it can still change your Q4 capacity position.
Note also that ASIN-level restock limits can cap individual SKUs even when your account looks acceptable overall. Your best seller can be throttled because of problems elsewhere in the catalog, which is a strong argument for clearing dead stock that appears to be costing you only storage.
Stranded inventory. Units sitting in fulfillment centers with no active listing damage your score while generating no revenue and accruing storage. Pull the Stranded Inventory report, relist or remove everything on it, and you have improved one of the four inputs in an afternoon with no forecasting judgment required.
Building the Q4 Forecast
Forecast in units, per SKU, per week. Aggregate revenue forecasts are comforting and useless for purchasing decisions.
Step one: corrected baseline
Take last year's Q4 units per SKU per week, then correct it for the things that distorted it:
- Add back stockout weeks. If you sold out on November 20, your recorded demand for that week is not demand, it is supply. Estimate what the trend implied.
- Remove one-off spikes from a deal or an external feature that will not repeat.
- Adjust for price changes between then and now.
- Drop discontinued and replaced SKUs rather than mapping them optimistically onto successors.
Step two: growth overlay
Apply your current year-over-year growth rate, measured over the trailing three months, not your aspiration and not your annual average. If you are up 22% over the last quarter, apply 22%, not the 40% in the business plan.
Step three: distribution across weeks
Q4 demand is not flat and it is not a single spike. It builds through October, peaks across Thanksgiving week into Cyber Monday, holds through the first three weeks of December, then collapses. Model at least four distinct periods rather than a monthly average, because a monthly average will leave you over-stocked in October and short in late November.
The reorder mechanics behind this are covered in our inventory reorder guide.
Promo and Deal Lift
The most over-estimated variable in Q4 forecasting, and the one that produces the most January dead stock.
How to model it honestly
- Use your own historical lift for the same deal type on the same or a comparable ASIN. Not a category benchmark and not the vendor's projection.
- Model a range, not a point. A low, expected and high case, then order against the expected case with the high case covered by safety stock rather than by committed inventory.
- Only apply lift where the deal is confirmed. Forecasting a lift for a deal you have submitted but not had accepted is how brands end up with 4,000 units and no promotion.
- Account for pull-forward. A large share of deal volume is demand that would have arrived later at full price. Net lift is smaller than gross lift, and the difference lands as a quiet slump in the following two weeks.
The pull-forward correction most models omit
If a deal sells 900 units in 48 hours and your normal weekly run rate is 200, the naive read is a 4.5x lift. But some of those buyers would have purchased in the following weeks anyway. Model the two weeks after a deal at a discount to your baseline, not at baseline, or you will forecast a trough as a shortfall and over-order to cover it.
Over-estimated deal lift produces excess inventory, which raises your utilization ratio, which can trigger a surcharge, while the units themselves age toward the next threshold and drag your IPI down. One optimistic assumption in September can touch four separate charges by February.
Safety Stock for Peak
Safety stock covers variability, not expected demand. Confusing the two is why brands carry too much of the wrong thing.
Setting the service level by margin, not by preference
Everyone wants 99%. The right answer depends on what a stockout costs versus what holding costs.
| Product profile | Service level | Reasoning |
|---|---|---|
| High margin, compact, hero SKU | 97–99% | Stockout cost dwarfs storage cost |
| Mid margin, standard size | 95% | Balanced |
| Thin margin, bulky | 90% | Storage and surcharge risk dominates |
| Long tail, slow mover | 85% or FBM | Not worth peak storage at all |
The Q4 adjustment
Lead time variance rises during peak because receiving slows as fulfillment centers prioritise outbound. That term is inside the formula, so the same service level requires more safety stock in Q4 than it does in June — which is a real cost and an argument for holding the buffer outside FBA where possible.
Want your fee stack modelled?
We will pull your storage and inventory age reports, calculate your utilization ratio, and show you which SKUs are about to cost you three charges at once.
Book a Strategy Call →The Ecom Profit Box
Eleven playbooks on listings, conversion, images, and email. Built for operators, no fluff, no email sequence.
Grab It Free →AWD vs FBA vs 3PL
| Dimension | FBA direct | AWD | Third-party 3PL |
|---|---|---|---|
| Q4 storage cost | Highest | Lower, off-peak rate extended | Usually lowest |
| Prime badge | Yes | Yes, once replenished | Only via MCF or SFP |
| Counts to capacity limits | Yes | No, held upstream | No |
| Speed to replenish | Already there | Automatic, Amazon-managed | Manual, you coordinate |
| Flexibility to other channels | Low | Low | High |
| Best for | Fast movers you will sell | Buffer stock for known winners | Overflow and multichannel |
The structural argument for holding buffer upstream
Safety stock is by definition inventory you hope not to sell. Holding it inside FBA during Q4 means paying peak rates on units whose entire purpose is to not move. Holding it upstream — in AWD or a 3PL — and replenishing on signal keeps the buffer cheap and keeps your FBA utilization ratio lower.
The tradeoff is replenishment lag. Buffer held outside FBA only works if you can move it in fast enough to matter, which argues for AWD's automated replenishment over a manual 3PL relationship when the destination is Amazon.
The honest caveat on AWD
Amazon's reported Q4 2025 results for AWD enrollees — more shipped units and materially fewer out-of-stock days — are Amazon's own figures for Amazon's own product. The mechanism is plausible and the extended off-peak storage rate is a real saving, but weight the numbers accordingly rather than treating them as independent evidence.
Our comparisons of AWD versus FBA inventory strategy and 3PL versus FBA versus self-fulfillment go deeper on both decisions.
The 120-Day Action Deadline
The single most useful operational rule in this article.
Whatever the current aged inventory threshold is, your action deadline is roughly 120 days of age. Not the threshold itself. By the time a unit hits the surcharge point, you have already lost the window in which cheap options existed.
Why 120
- Discounting takes weeks to clear stock, and deeper discounts take longer to decide and approve than anyone plans for.
- Removal orders take time to process, and there is only a short grace period after a surcharge is assessed to create one.
- Switching a SKU to FBM to eliminate storage entirely requires operational setup.
- Liquidation partners need lead time.
The comparison to run at 120 days
Removal costs roughly one to two dollars per unit depending on size and weight. Compare that against the storage and surcharges you will pay across the remaining months you expect to hold it. For slow movers heading into Q4, removal is frequently cheaper than holding, and that calculation is not close.
Put a 90-day review and a 120-day decision point into your inventory process. Reviewing monthly and acting at 181 means every decision is made after the cheap options have expired. This is a calendar problem more than an analytical one.
If You Are Already Over-Stocked
Triage in cost order, cheapest lever first.
- Pull the Inventory Age report and sort by cubic footage, not by unit count. Your problem is volume.
- Fix stranded inventory first. Free, immediate, improves IPI and stops paying storage on units that cannot sell.
- Identify the bulky slow movers. A small number of ASINs almost always account for most of the exposure.
- Discount before Q4 rates begin, not during. A markdown in September clears at $0.78 per cubic foot; the same markdown in November has already paid peak rates for two months.
- Compare removal against holding cost for anything you will not clear by December.
- Move genuinely slow SKUs to FBM to eliminate storage entirely.
- Stop replenishing anything above 18 weeks of cover until the ratio comes down.
What not to do
Do not liquidate your fast movers to bring the ratio down. The ratio is volume-weighted, so cutting the SKUs that actually turn hurts your sell-through input and your revenue while barely touching the volume driving the charge. Cut the bulky slow movers, which are almost always the smaller revenue contributors anyway.
And do not solve a Q4 storage problem by cancelling inbound shipments of products that sell. Being out of stock on a winner during peak costs more than any storage line on the bill.
The January Drawdown Plan
Written in September, executed in December, because January is too late to plan it.
The January problem
Peak fulfillment fees run into mid-January. Q4 returns arrive and re-enter inventory. Demand collapses. Whatever did not sell is now both older and sitting in a month with elevated per-unit costs. It is the worst combination in the calendar and it is entirely predictable in advance.
Decide these now
- Your clearance trigger. A specific cover threshold that automatically initiates markdown, decided before you are emotionally attached to the margin.
- Your removal list. Which SKUs get removed rather than discounted, and at what age.
- Returns handling capacity for the surge beginning December 26, since returned units re-entering inventory affect both your volume and your age profile.
- Your Q5 pricing. Late December into mid-January carries genuinely cheap traffic from gift card redemption against collapsed competition. That is the best clearance window of the year, and it only works if the stock is priced to move before storage compounds.
- Your first restock date for the products that sold through, so you are not out of stock in a quiet month while competitors restock.
Contribution margin per unit with peak fulfillment fees, all storage charges and the actual returns rate applied — not Q4 revenue. Since peak fees run to mid-January and returns settle through January, that number is not readable until February. Book the review then, and make next year's ordering decisions from it rather than from the revenue headline.
The financial modelling is covered in our contribution margin playbook and the 13-week cash flow model, and the full fee detail in the 2026 FBA fee breakdown.
The Short Version
- Three charges stack on the same unit: peak base storage, aged inventory surcharge and storage utilization surcharge. None cancels another, and a bulky slow-moving SKU in December can incur all three.
- Standard-size storage runs $0.78 per cubic foot January to September and $2.40 October to December. Fees are assessed on the 15th from the prior month's daily average cubic footage, so December's bill reflects November's decisions.
- The utilization surcharge applies above 25 cubic feet daily volume and a utilization ratio over 22 weeks. Start slowing replenishment at 18 weeks, because the ratio is a trailing 13-week average that moves slowly.
- Your action deadline is roughly 120 days of inventory age, not the surcharge threshold. By 181 days the cheap options have expired.
- Forecast in units per SKU per week with stockout weeks added back, current trailing growth applied, and at least four distinct Q4 periods modelled.
- Model deal lift as a range, only where confirmed, and discount the two weeks after a promotion for pull-forward.
- Safety stock covers variability, not demand. Set the service level by margin and bulk, and hold buffer upstream in AWD or a 3PL rather than paying peak rates on units meant not to move.