Amazon Inventory Forecasting: The Operator’s Guide for FBA Sellers (2026)
A plain-English guide to forecasting Amazon inventory — from what the metrics actually mean, to how the 2026 fee changes affect your math, to the honest trade-offs between spreadsheets and software. Written by a former 3PL operator, not a SaaS marketing team.
Quick Answer
Amazon inventory forecasting is the process of predicting how much stock to reorder, when to reorder it, and how much to send to FBA — using sales velocity, lead times, and safety stock. For most FBA sellers, accurate forecasting is the single biggest lever for avoiding stockouts, aged inventory surcharges, low-inventory fees, and lost Amazon search ranking.
The core math is simple: (Average Daily Sales × Lead Time) + Safety Stock = Reorder Point. The complexity isn’t the formula — it’s getting accurate lead times, recognizing when safety stock rules stop working, and adjusting to Amazon’s 2026 fee structure, which punishes both overstocking and understocking harder than it used to.
What Amazon inventory forecasting actually is
Forecasting on Amazon isn’t about predicting the future. It’s about answering three concrete operational questions with math instead of gut feel:
What should I order this week?
From the manufacturer. In what quantity. Based on how fast this SKU is selling across every channel, not just Amazon.
When do I need to trigger it?
The reorder point — the stock level that tells you “place the PO now” so new inventory arrives before the current stock runs out.
How much do I send to FBA?
Not all inventory goes to FBA at once. How much stays at your warehouse or AWD, and how much ships into FBA’s sellable pool right now.
If your current inventory process — whether that’s a spreadsheet, a piece of software, or a gut-feel restock meeting — answers these three questions consistently and correctly, forecasting is working. If it doesn’t, you’re bleeding margin somewhere. The fees in 2026 make that bleed bigger than it used to be.
The five metrics every FBA forecast runs on
You don’t need to master every inventory textbook. Five metrics do 90% of the work. Understand these and you understand forecasting:
Units sold per day, per SKU, per channel. Usually measured over 30 or 90 days. Velocity = Total Units Sold ÷ Days in Period. This is the foundation of every other number.
Days from placing a PO to having sellable inventory at FBA. Includes factory production + freight + customs + FBA receiving. For overseas sellers this is typically 60–90 days, not the 21 days your supplier quotes.
How many days your current inventory will last at current velocity. DOS = Current Stock ÷ Daily Velocity. This is what Amazon uses to calculate your restock limits and low-inventory fees.
Buffer inventory to protect against demand spikes and late shipments. For most FBA sellers in 2026, aim for 30–45 days (enough to avoid the low-inventory fee at 28 days, not so much that you trigger the aged-inventory surcharge at 181 days).
The stock level that triggers a new PO. Reorder Point = (Daily Velocity × Lead Time) + Safety Stock. When on-hand inventory drops to this number, place the order. Full guide + calculator.
How the 2026 Amazon fee changes change the math
Amazon stacked three fee changes on FBA sellers in the first half of 2026:
- A 3.5% fuel and logistics surcharge on every FBA fulfillment fee (effective April 17, 2026)
- The aged inventory surcharge threshold dropped from 271 days to 181 days — 90 days earlier than before
- The low-inventory-level fee is now calculated per FNSKU, not per parent ASIN
The practical effect for forecasting: your safe inventory window got smaller on both sides. Hold too much and the aged surcharge hits earlier. Hold too little (under 28 days) and the low-inventory fee hits per variation, not just per product. Every forecasting model built before 2026 needs recalibration — tighter safety stock, shorter reorder cycles, per-FNSKU reorder points, and more use of AWD as a buffer layer instead of overstocking FBA.
Full breakdown of the fee changes and specific adjustments: Amazon FBA Inventory Forecasting in 2026. For AWD tracking specifically: How to Track Amazon AWD Inventory in 2026.
Spreadsheet vs software: when each one is the right tool
Every FBA seller goes through the same progression: gut feel → spreadsheet → software. The question isn’t which tool is better — it’s which tool fits where you are right now.
| Your situation | Best tool | Why |
|---|---|---|
| Under 50 SKUs, Amazon-only, stable demand | Gut feel + Seller Central reports | Simple catalog, predictable velocity, reorder decisions are obvious. |
| 50–200 SKUs, Amazon + maybe Shopify | Free Excel template | Days-of-coverage model with FBA/AWD/warehouse split. 45 min a week. Free. |
| 200+ SKUs, multiple channels, variations | Dedicated forecasting software | Spreadsheet can’t keep up with per-FNSKU reorder points, velocity refreshes, and multi-channel demand signals. |
| Growing fast, stockouts happening | Software + managed service | You don’t have time to learn it and run it. Let a human analyst handle the weekly restock plan. |
| $5M+ in revenue, enterprise needs | ERP or specialized platform | Forecasting becomes one module in a larger financial and operational system. |
The honest take: most FBA sellers doing $500K–$5M in revenue are in the sweet spot where dedicated forecasting software earns its keep — not because spreadsheets can’t handle the math, but because keeping a spreadsheet accurate every week across 200+ SKUs becomes a part-time job. Software eats that job. The cost math is usually clear once you run it.
Six common Amazon forecasting mistakes (and how to fix them)
I’ve watched these same mistakes play out at dozens of brands. They’re not “gotcha” mistakes — they’re the kind of reasonable decisions that make sense until you see the data six months later:
1. Using the supplier’s quoted lead time instead of real lead time
The factory says “21 days production.” You use 21. But your real lead time also includes ocean freight (28 days), customs (3–7 days), and FBA receiving (7–21 days). Your actual lead time to sellable FBA stock is 60–80 days, and that’s the number that goes into your reorder point formula.
2. Calculating reorder points at the parent ASIN, not per FNSKU
With the 2026 low-inventory-level fee now calculated per individual FNSKU, a single reorder point across all variations isn’t granular enough. Every size, color, and bundle needs its own reorder trigger. A well-stocked Medium hiding a stocked-out Small will now cost you per-FNSKU.
3. Forecasting from Amazon velocity only when you also sell on Shopify
If the same SKU sells on both channels, your total demand is the sum, not just the FBA number. Shopify-native forecasting tools that don’t see Amazon, and Amazon-only tools that don’t see Shopify, both under-forecast for multi-channel sellers. How to forecast across both channels.
4. Sending too much inventory to FBA at once
The aged-inventory surcharge starts at 181 days in 2026. If you send 200 days of supply in one shipment, the last 19 days will pay the surcharge. Smaller, more frequent shipments — using AWD or your own warehouse as the buffer — avoid this entirely.
5. Holding too little safety stock to avoid storage fees
Overcorrecting from mistake #4. If safety stock drops below 28 days, the low-inventory fee triggers — which costs more than modest storage fees. The sweet spot is 30–45 days of safety stock for most SKUs.
6. Forecasting new products the same as mature SKUs
A SKU with 30 days of sales history has no reliable average. Statistical forecasting returns nonsense. Hold extra safety stock on new launches, review weekly, and don’t trust the math until you have 90+ days of real velocity data.
What forecasting won’t solve
Forecasting is a tool, not a crystal ball. It won’t save you from a listing suppression, a review bomb, a competitor entering with a 40% lower price, or Amazon’s algorithm deciding your product is no longer relevant. Those are market events — forecasting only tells you how to react to them efficiently once they happen.
What good forecasting will do: keep you from being the seller who runs out of their best SKU right before Prime Day, or the one who pays aged-inventory surcharges on Q4 leftovers through March. It raises the floor. The ceiling is still up to everything else about your business.
The 30-day Amazon forecasting setup
If you’re starting from gut feel and want to move to a real system, here’s the minimum-viable 30-day plan:
Week-by-week
(Daily Velocity × Lead Time) + Safety Stock for every SKU. Use 30–45 days of safety stock for most. Flag anything below its reorder point.
The free forecasting template has this structure built in — just drop in your SKU data and it runs the math. By day 30 you’ll have the first month of real forecast-driven ordering behind you, and the first signal of what’s working and what isn’t.
Frequently asked questions
What is Amazon inventory forecasting?
Amazon inventory forecasting is the process of predicting how much stock to reorder, when to reorder it, and how much to send to FBA, based on sales velocity, lead times, and safety stock. Accurate forecasting helps FBA sellers avoid stockouts, aged inventory surcharges, low-inventory fees, and lost Amazon search ranking from going out of stock.
What’s the formula for Amazon reorder points?
The standard reorder point formula is: (Average Daily Sales × Lead Time in Days) + Safety Stock. For Amazon FBA, lead time must include factory production, ocean freight, customs, and FBA receiving time — not just the supplier’s quoted production days. A SKU selling 10 units/day with a 63-day real lead time and 14 days of safety stock has a reorder point of 770 units.
How much safety stock should Amazon FBA sellers keep?
For most FBA sellers in 2026, aim for 30–45 days of safety stock. This stays above the 28-day threshold that triggers the low-inventory-level fee and well below the 181-day threshold that triggers the aged-inventory surcharge. Adjust up for volatile SKUs, seasonal products, or new launches.
How do the 2026 Amazon fee changes affect forecasting?
Three fee changes in 2026 tightened the safe inventory window: the 3.5% fuel surcharge on fulfillment fees, the aged inventory surcharge threshold dropping from 271 to 181 days, and the low-inventory-level fee now calculated per FNSKU. Sellers need tighter safety stock (30–45 days vs the old 45–60), shorter reorder cycles, per-FNSKU reorder points, and more use of AWD as a buffer.
Can I forecast Amazon inventory in a spreadsheet?
Yes, up to about 150–250 SKUs on a single channel. A well-built spreadsheet using a days-of-coverage model can handle the core forecasting math for small and mid-catalog brands. Above that, spreadsheets struggle with per-FNSKU reorder points, daily velocity refreshes, and multi-channel demand signals — that’s where dedicated software starts earning its keep.
What’s the difference between FBA sellable inventory and total Amazon inventory?
FBA sellable inventory is the stock that’s checked in at a fulfillment center and available to ship to customers. Total Amazon inventory also includes FBA inbound (in transit, not yet received, usually 7–21 days in receiving) and AWD stock (upstream buffer in Amazon’s Warehousing & Distribution network). Accurate forecasting tracks all three layers separately.
How often should I update my Amazon inventory forecast?
For most brands, weekly refreshes are the right cadence. Daily velocity changes are usually noise; monthly is too slow to catch demand shifts. A weekly Monday review of velocities, on-hand stock, and open POs across every SKU is the operational rhythm that matches Amazon’s restock cycle and FBA’s receiving timelines.
Does Amazon provide inventory forecasting tools for sellers?
Amazon provides some forecasting signals in Seller Central (restock recommendations, forecasted demand in the Inventory Dashboard), but they’re channel-limited (Amazon only), SKU-level without variation-specific reorder math, and don’t integrate with non-Amazon channels. For multi-channel sellers or anyone managing 100+ SKUs, third-party forecasting software is typically required to get accurate per-SKU reorder points and multi-layer inventory visibility.