The Complete FBA Inventory Management Guide
Demand forecasting, lead times, safety stock, reorder points, capacity limits, IPI score, and the operational cadence for managing FBA inventory at scale. Built by an 8-figure seller.
Inventory management is the operational discipline that determines whether an Amazon business is profitable or just busy. Get it right and your capital recycles fast, your storage fees stay low, and your stockouts are rare. Get it wrong and you bleed cash to long-term storage fees on slow movers, miss revenue on stockouts of fast movers, and watch your IPI score drag down your capacity limits.
This guide walks through the full operational stack: how to measure velocity correctly, how to derive safety stock from demand variance instead of guessing, how to compute reorder points that account for ACTUAL lead times, how to manage capacity limits and restock limits, and how to build the weekly cadence that keeps everything running. By the end you will know what to look at, when to look at it, and what to do when each metric goes sideways.
Sales velocity: the foundation everything else depends on
Every inventory decision starts with sales velocity — how many units of a SKU you sell per day, per week, or per month. Velocity is the input to forecasts, reorder points, safety stock, days of cover, and storage cost projections. If velocity is wrong, every downstream number is wrong.
The trap with velocity: which window do you measure? The 30-day average smooths short-term noise but lags trend changes. The 7-day average tracks recent shifts but is volatile. Neither is "right" by itself; the right answer is to track multiple windows side by side and pay attention to the direction of change.
A SKU with 30-day velocity of 8 units/day and 7-day velocity of 14 units/day is accelerating. If you set your reorder point based on the 30-day number, you will be late on the next order by the time the 7-day catches up to the 14-day reality. A SKU with 30-day velocity of 8 units/day and 7-day velocity of 3 units/day is decelerating — possibly seasonal, possibly a quality issue, possibly a competitor undercutting. Either way, your forecasts based on the 30-day number are about to over-order.
Seasonal SKUs compound the problem. A Christmas item has Q4 velocity 10x its baseline. The annual average hides that completely. Good inventory tools weight velocity by recency AND adjust for seasonality. SellerVault's Monte Carlo forecasting layers seasonal indices on top of recency-weighted base velocity to produce per-SKU probability distributions instead of point estimates.
Days of cover: the most useful inventory metric
Days of Cover (DOC) is the number of days your current on-hand inventory will last given recent sales velocity. The formula is straightforward: DOC = on-hand units / average daily sales.
Why DOC is the most useful single inventory metric: it abstracts away absolute unit counts and converts everything to a single time-based number that is directly comparable across SKUs. A SKU with 1,000 units of cover can wait — it has plenty of runway. A SKU with 12 days of cover and a 14-day lead time is already late on the reorder. DOC turns the question "do I have enough?" into "do I have enough TIME?" — which is the question that actually drives decisions.
The right restock workflow ranks SKUs by DOC ascending and processes the lowest-DOC items first, factoring in lead time and safety stock requirements before generating order quantities. Spreadsheet-based restock workflows usually skip DOC entirely and rank by absolute unit count instead, which produces wrong priorities — you end up restocking SKUs with 200 units of cover before SKUs with 30 days of cover because the 200-unit SKU "looks low" in absolute terms.
Lead time: stop using the supplier's number
Lead Time is the total elapsed time from "I place a purchase order" to "the inventory is sellable in Amazon's FBA network." It includes supplier production, freight from supplier to your warehouse or prep center, prep work, inbound shipment to Amazon, Amazon's receive time, and the gap between received and sellable status.
Most operations underestimate lead time by 30-50% because they only count the supplier production component. A real lead time breakdown for a typical wholesale order:
| Phase | Days |
|---|---|
| Supplier production | 7 |
| Freight to your warehouse | 4 |
| Prep work | 2 |
| Inbound shipping to Amazon | 5 |
| Amazon receiving | 3 |
| Total | 21 |
If your DOC drops to 21 days you are ALREADY late on the reorder — there is no slack for normal variance, much less for the inevitable supplier delays.
The right approach is to track ACTUAL lead time per supplier per SKU, not the supplier's claimed lead time. The two are usually different. SellerVault's restock engine measures actual lead time from the time gap between PO creation and "received" status across past orders, then surfaces a confidence interval — useful for flagging suppliers whose lead time has been creeping up.
Reorder point and safety stock
Reorder Point (ROP) is the inventory level at which you should trigger the next supplier order. The formula is ROP = (average daily sales × lead time in days) + safety stock. When on-hand falls to ROP, the math says you should place an order today.
Worked example: a SKU sells an average of 8 units per day, has a 21-day lead time, and you want 10 days of safety stock. ROP = (8 × 21) + (8 × 10) = 168 + 80 = 248 units. When on-hand drops to 248 units, place the order.
Safety stock is the buffer you hold above expected demand to absorb two kinds of variance: demand variance (some weeks sell more than the average) and lead time variance (some shipments arrive later than expected). Without safety stock, even perfect average forecasts produce stockouts about half the time, because half your weeks are above average.
The simple safety-stock formula is safety stock = Z × σ × √(lead time in days) where Z is the desired service level (1.65 for 95% in-stock probability, 2.33 for 99%) and σ is the standard deviation of daily demand. In practice most operators do not bother with the formula and instead pick a "safety days" number by feel — 7 days for stable SKUs, 14-21 for volatile ones. SellerVault's Monte Carlo forecasts produce probabilistic distributions per SKU so you can pick a target service level (95%, 99%) and the engine derives the right safety stock from actual demand variance instead of using a flat buffer across the catalog.
Capacity limits and restock limits: two different constraints
These two get conflated constantly. They are different.
Capacity Limits cap how much you can STORE in FBA at any given moment, measured in cubic feet per storage type (standard, oversized, apparel, footwear, etc.).
Restock Limits cap how much you can SEND to FBA in a given month, measured in units per storage type.
A common gotcha: you have 5,000 cubic feet of remaining capacity but only 200 units of remaining standard restock for the month. You CANNOT use that capacity until next month, no matter how empty your bins are.
Both limits are set by Amazon based on a combination of your IPI score, sales velocity, and Amazon's overall warehouse capacity. They are reset monthly. You can purchase additional capacity through the Capacity Manager auction, which is occasionally cost-effective for Q4 ramp but usually too expensive to bother with.
The defensive playbook:
- Check both limits every Monday before restock planning.
- Prioritize fast movers if you are restock-limited.
- Use AWD to sidestep both limits for SKUs that do not need to be in FBA right now.
- Improve IPI to raise both limits next month.
SellerVault's restock planner shows remaining capacity AND remaining restock limit side by side, and warns when a recommendation would exceed either. Most spreadsheet workflows ignore both and discover the constraint only when an inbound shipment gets rejected at creation.
IPI score: the four levers
IPI (Inventory Performance Index) is the 0-1000 score Amazon uses to evaluate how well you manage FBA inventory. It directly affects your monthly capacity and restock limits. A high IPI gets you generous limits; a low IPI restricts your operations and may trigger storage overage fees.
IPI is derived from four components:
- Excess Inventory Percentage — the share of your SKUs that have more than 90 days of cover. To improve: clear excess via removal orders or aggressive markdowns.
- Sell-Through Rate — units sold over the last 90 days divided by average inventory over the same window. To improve: increase velocity OR reduce inventory levels.
- Stranded Inventory Percentage — share of SKUs that have FBA inventory but cannot be sold (suppressed listings, gated brands, etc.). To improve: clean up stranded inventory weekly via the Stranded Inventory dashboard.
- In-Stock Rate — share of high-demand SKUs that have not stocked out. To improve: better restock cadence on the top 20% of SKUs by velocity.
These four are the only levers that move IPI. Generic "improve inventory health" advice is useless because IPI does not measure generic health. It measures these four ratios.
Aged inventory: the silent profit killer
Amazon charges escalating storage fees on inventory that has been in FBA for too long. The Aged Inventory Surcharge kicks in at 181 days, escalates at 271 days, escalates again at 365 days, and the Long-Term Storage Fee adds another layer at the 365-day threshold. For a slow-moving SKU, storage cost can easily exceed the unit's gross profit margin within 12 months — turning every "I will sell it eventually" decision into a slow loss.
The defensive playbook is simple but requires discipline:
- Pull the FBA Inventory Age Report monthly.
- Identify any SKUs with units approaching the 180-day window.
- For each, decide: aggressive markdown to accelerate velocity, OR removal order back to your warehouse, OR removal order disposal.
- Execute the decision the same week. Do not let it slide.
The trap is putting these decisions off because they feel small individually. They compound. SellerVault surfaces aged-inventory cost projections on the inventory dashboard so you can see the future fee impact of decisions you make today.
The weekly inventory operations cadence
Inventory management at scale fails when sellers treat it as a "look at it when something goes wrong" task. By the time something is visibly wrong, it is too late to fix. The right approach is a fixed weekly cadence that touches every metric whether or not anything is on fire.
A working weekly cadence:
Monday morning (30 minutes):
- Check capacity limits and restock limits for the month.
- Run the restock planner. Process the lowest-DOC SKUs first.
- File POs for anything below reorder point.
- Check the Stranded Inventory dashboard. Fix anything fixable; remove anything not.
Wednesday (15 minutes):
- Mid-week velocity check. Are any SKUs accelerating or decelerating unexpectedly?
- Adjust safety stock targets on any SKUs with major velocity shifts.
Friday (30 minutes):
- Weekly reimbursement audit (warehouse lost/damaged categories).
- Review any inbound shipments delivered this week — verify received quantities match shipped.
- Pull aged-inventory report once a month, not weekly. Once a month is enough for the monthly reset cadence.
That is roughly 75 minutes per week for a meaningful FBA operation. The investment pays for itself many times over.
Frequently asked questions
How often should I restock?
Weekly is the sweet spot for most catalogs. Daily restock cycles do not move enough inventory to justify the operational overhead; monthly cycles let things slip too long. Set Monday morning as your fixed restock review day so it does not fall off the schedule.
What is a good IPI score?
Amazon does not publish a hard threshold but anything above 500 unlocks generous capacity. Below 400 you start seeing capacity restrictions. Above 700 you are in the top tier of FBA sellers operationally. The score updates weekly so improvements compound fast.
Should I use AWD or send directly to FBA?
Depends on the inbound placement fee math. AWD has higher per-unit storage cost but bypasses the inbound placement fee structure entirely. For high-volume SKUs that would otherwise hit IPSF on every restock cycle, AWD is usually cheaper. For low-volume SKUs, direct FBA is cheaper. SellerVault's placement intelligence calculates this automatically per shipment.
How do I handle seasonal SKUs in forecasting?
Build a seasonal index from at least one full year of history (ideally two), and apply the index as a multiplier on top of base velocity. SellerVault's Monte Carlo forecaster does this automatically. Manually, you can compute seasonal indices in a spreadsheet by averaging each month's sales as a percentage of the annual average.
What is the difference between Days of Cover and Days of Stock?
They are the same thing — different vendors use different terminology. Both mean "how many days will my current on-hand inventory last at recent velocity." SellerVault uses "Days of Cover" because it is the more common term in operations literature.