Amazon Inventory Management: What It Actually Means and How the Math Works

What Amazon inventory management actually is

Amazon inventory management is the discipline of knowing what stock you have, where it sits, and when to reorder it, so a listing never goes out of stock and cash never sits frozen in units nobody's buying. For a seller, that means tracking sales velocity, lead times, and storage costs inside Seller Central and turning that into reorder decisions before it turns into a stockout or a storage fee.

The phrase gets used two different ways, and mixing them up wastes time. "Inventory management at Amazon" usually means Amazon's own internal logistics — the robotics, the fulfillment-center routing, the automation that gets a package to a doorstep in two days. "Inventory management for Amazon" (or "of Amazon," or just "amazon inventory management" on its own) almost always means something narrower: a third-party seller managing their own stock on Amazon's marketplace. This page is about the second one — what a seller controls, not what Amazon's warehouses do.

If you sell through Fulfillment by Amazon (FBA), Amazon stores and ships the units. It does not decide how many to make, when to reorder, or what price to sell at. Fulfilled by Merchant (FBM) hands you both storage and shipping yourself. Either way, forecasting and reorder timing stays on you — Amazon gives you data and a couple of built-in tools, not a decision-maker.

The mechanics: reorder point, safety stock, and a worked example

Every inventory system, regardless of software, runs on the same core calculation: reorder point equals average daily sales times lead time, plus a safety-stock buffer for the unexpected.

Say a SKU sells 50 units a day, and it takes 21 days from placing a purchase order to units being sellable in an Amazon warehouse. The baseline reorder point is 50 × 21 = 1,050 units — that's the demand you'll burn through before a fresh order can possibly land. Add a safety-stock cushion of 10 days for supplier delays or a sales spike: another 500 units. Reorder point: 1,550 units. When on-hand plus inbound inventory drops to that number, the purchase order should already be placed, not started.

The part almost every guide skips: that 1,550 figure is only correct until something changes the inputs. A coupon, a PPC push, or a competitor going out of stock can double daily velocity overnight, which moves the reorder point without warning. Across the $500M+ in managed revenue Full Circle has managed for 100+ brands, the same pattern shows up repeatedly — the stockout usually wasn't a forecasting failure, it was a reorder point calculated once and never updated against what the ad account was about to do to demand.

The tools: Seller Central, FBA, ASCS, and where a full system fits

Amazon's own inventory management system inside Seller Central covers the basics: current stock levels, a restock recommendation based on trailing sales velocity, the Inventory Performance Index (IPI) score, and Amazon Supply Chain Services (ASCS) if you want Amazon involved further upstream. These are real tools, and they work fine for straightforward, single-marketplace sellers with the time to check them.

They have real limits too. The restock recommendation is backward-looking — it has no idea a promotion or an ad campaign is about to change velocity. IPI is a 12-week rolling average, which means a bad month gets diluted by two good ones and a genuinely bad quarter takes a while to show up as a storage-limit cut. Neither tool touches reimbursements for lost or damaged units, dimensional-weight fee errors, or the removal-versus-liquidation math on dead stock.

Worth being precise about category here, because the tools aren't interchangeable. A 3PL physically stores and ships product — Amazon's FBA network is one option, a private 3PL is another, and neither is optional to have if you're not using the other. An ERP or multichannel inventory system runs the whole operation: purchasing, accounting, multi-channel stock sync. A brand already running one of those shouldn't expect an Amazon-specific tool to replace it. A research suite is built to find products worth selling, not to work a live fee dispute or a stuck reimbursement. None of these do what the others do, and paying for three tools that overlap on nothing important is its own kind of leak.

The methods worth knowing

Once the reorder math is in place, the method you layer on top depends on the SKU. None of these are Amazon-specific — they predate e-commerce by decades — but they come up constantly in how sellers talk about managing stock on the platform. A brand with a handful of SKUs, like Ridge, manages this differently than a brand with a wide size-and-color matrix, like HexClad — more SKUs means ABC analysis matters more, not less.

The mistake almost everyone makes — including us

The most common mistake isn't a missing spreadsheet. It's treating Amazon's restock recommendation as a decision instead of an input. The number is built entirely from what already sold; it has no visibility into what's planned. We've made this mistake too — approved a suggested reorder quantity without checking whether a deal or a PPC ramp was scheduled for the same window, and watched the SKU go out of stock two weeks into a campaign it was supposed to support.

The fix isn't complicated, it's just a habit: check the promotion and ad calendar against the reorder point before approving the order, not after the stockout. If the leak you're chasing turns out to be in the ad account — wasted spend, a listing that converts poorly, bids that don't match the plan — that's a different problem with a different fix, and it belongs with Dr. PPC rather than an inventory review.

When the number is already bad — what to do next

Sometimes you catch this after the fact: the IPI score already dropped, the reorder already landed late, or the last "fix" didn't stop the stockout. Each has a different next step, and panicking into the wrong one usually costs more than the original problem.

• IPI score already low: don't liquidate broadly to fix the average. Work the specific SKUs dragging it down — aged units past the surcharge threshold are usually the biggest single lever, and the removal-versus-liquidation decision should be made SKU by SKU, not catalog-wide.
• Reorder already late: expediting freight is expensive but often cheaper than the lost rank from a multi-week gap, especially mid-campaign. If expediting isn't viable, throttle ad spend to stretch remaining stock rather than let it run out mid-flight.
• The last fix didn't work: check whether an input changed again — lead time, supplier minimums, a second demand spike — before assuming the process itself failed. Most "the forecast was wrong" complaints are actually "the forecast was never updated."

Where Dr. Stock fits

Dr. Stock is Amazon inventory and supply chain run as a managed product, not a piece of software you configure yourself. It goes after money already leaking — stockouts and the reorder timing behind them, cash trapped in slow-moving SKUs, storage and aged-inventory fees, FBA fee errors, and reimbursement recovery — and checks it against real numbers rather than promising a result in advance. Purchasing decisions always go to a human, regardless of what autonomy setting a client chooses elsewhere. Whether or not that's the right fit for a given seller, the reorder-point math and the IPI mechanics above hold either way — that part doesn't change based on who's running it.