Inventory & Fulfilment

Why Inventory Is the Second Pillar

Going out of stock is the most expensive thing that can happen to an Amazon product, and it is worse than the lost revenue suggests. You do not simply pause. Your competitors take the sales you would have made — and with them the rank, the reviews and the advertising budget that those sales would have funded. You come back to a worse position than you left. Watch the walkthrough Prefer to watch? One hundred and sixteen seconds on what a stockout actually costs, the two halves of inventory management, and the standard worth holding yourself to. Why it outranks everything else Every other improvement is conditional on being in stock. Better advertising does nothing when there is nothing to sell. A stronger main image does nothing. Improved organic rank does nothing, and worse, it decays while you are unavailable. So if you have gone out of stock repeatedly in the last year, inventory is where your attention belongs — ahead of advertising, ahead of listing work, ahead of new products. For a brand with a recurring stockout problem, simply fixing it is often worth more than every other optimisation combined. Stock levels, per ASIN The two halves of the job Inventory management is not complicated. It is a small number of things that must be done reliably. Monitoring — watching stock against sales weekly, so you see a stockout coming with enough time to act. Planning — deciding purchase orders so the stockout never becomes likely in the first place. Most sellers do some of the first and almost none of the second, which is why stockouts feel like emergencies rather than forecasting errors. What monitoring actually means Watch inventory levels against sales velocity, weekly. The number that matters is not units in stock. It is days of inventory — units divided by current daily sales — because that reframes stock as time, which is the thing you actually need. And watch for velocity changes in both directions. If you planned for 50 units a day and you are now selling 75, your runway just shortened by a third and nothing on the stock screen looks different. If you are selling 30, you have money sitting still and a storage bill. The weekly cadence exists so that change gets noticed while there is still time to respond. A monthly check on a product with six weeks of cover is not monitoring, it is luck. Where the data comes from The traditional answer is to wire a spreadsheet to your account through a third-party service and maintain a live dashboard of available, reserved and inbound units per child ASIN. TrackIQ reports that directly. Available, reserved and inbound units, days of inventory, and sales velocity sit together per ASIN, which is the whole of the monitoring job without the spreadsheet in between. See How to Monitor Your Stock Levels and Catch a Stockout Early. And it can tell you rather than wait to be asked. A low-stock alert on your top products removes the failure mode where the weekly check gets skipped in a busy week — see Setting Up Smart Alerts. The goals of the process Three, in order: 1. Spot ASINs heading for a stockout, early enough to act. Weeks of warning, not days. 2. When one is unavoidable, manage it deliberately rather than simply running out. There is a considerable difference in profit between the two — see When You Are Going Out of Stock. 3. After a stockout, get back to normal sales as fast as possible. Rank decays while you are unavailable, and recovering it takes deliberate effort rather than time. What good looks like 98%+ in stock across your active catalogue. That is the standard a relaunch assumes, and it is the standard advertising assumes too — driving velocity into a product that will run out is worse than not driving it at all. Never stocking out on a top seller. At all. Your best products earn a different level of care and a larger buffer, because a stockout there costs more than a stockout everywhere else combined. Related - Monitoring Inventory Against Sales — the weekly check, in detail. - When You Are Going Out of Stock — turning an unavoidable stockout into a profitable one. - Inventory Forecasting and Purchase Orders — making stockouts unlikely in the first place.

Monitoring Inventory Against Sales

The weekly inventory review is a short job with one purpose: find the ASINs whose runway has changed before that change becomes a stockout. It takes minutes when the data is in front of you, and it is the single highest-return recurring task in the business. Watch the walkthrough Prefer to watch? One hundred and five seconds on why runway is a ratio rather than a number, the four figures worth reading, and the failure mode that is not looking at all. Think in days, not units Units in stock is not a useful number on its own. Days of inventory is, because it answers the only question that matters: how long until this runs out at the rate it is currently selling? Days of inventory = units available ÷ current daily sales velocity A product with 4,000 units looks healthy until you notice it sells 200 a day and the next shipment lands in 35 days. Stock and days of cover, per ASIN What to look at Per child ASIN, four numbers: Available units — sellable right now. Reserved units — committed to orders or in transfer, and not available to sell. Counting these as stock is a common way to be surprised. Inbound units — on their way to Amazon, with a date attached. Units without a credible arrival date are not cover. Sales velocity — recent average daily units. Recent matters; a 90-day average hides a product that has doubled in the last fortnight. Then the derived views: days of inventory at Amazon, days including your warehouse or 3PL, and days across the whole pipeline. They answer different questions — the first is when you stop selling, the last is whether you have a supply problem or a logistics problem. Watch the velocity, not just the level This is the part that gets missed. The stock number can be unchanged week to week while your runway collapses, because runway is a ratio. Selling faster than planned — planned for 50 a day, now doing 75 — and your cover has dropped by a third with no change to the stock figure. Good news that becomes a stockout if nobody notices. Selling slower than planned — planned 50, doing 30 — and you are now carrying more stock than the product justifies. That is cash sitting still, plus storage fees, plus long-term storage surcharges if it drags. So compare this week's velocity to last week's, and to what you assumed when you ordered. A change of 30% or more in either direction needs a decision. The weekly routine Do it at a fixed point in the week, alongside your other numbers rather than as a separate task. It only works if it is habitual. 1. Sort by days of inventory, ascending. The problems are at the top. 2. For anything under your lead time plus a buffer, act. If replenishment takes 45 days and an ASIN has 40 days of cover, that is already a stockout unless something changes. 3. Check velocity changes across the catalogue. Look for anything that moved sharply, in either direction. 4. Flag your top sellers separately regardless. They deserve a look even when the numbers look fine, because the cost of being wrong is much higher. 5. Note anything that came back into stock, and push it back to normal sales deliberately — recovery is not automatic. Acting on what you find Heading for a stockout, with time to fix it — expedite the next shipment, move units from your warehouse or 3PL, or place an emergency order. Heading for a stockout with no time — slow sales down deliberately and take more margin per unit while it lasts. See When You Are Going Out of Stock. Overstocked — the levers are the reverse: promotions, coupons, deals, and more advertising spend. An overstocked product is a candidate for the advertising budget you were about to spend elsewhere. Velocity up sharply — revise the forecast now, not at the next planning cycle. The purchase order you are about to place was sized on the old number. Making it harder to skip The honest failure mode is not misreading the data, it is not looking during a busy week. Set alerts on the products that matter — low days-of-cover on your top sellers, and sharp velocity changes anywhere. That turns the weekly review from the only line of defence into a second one. See Setting Up Smart Alerts and Alert Categories Explained. And keep it in one place. The reason this task gets skipped is usually that assembling the view takes longer than reading it. See How to Monitor Your Stock Levels. Common mistakes Counting reserved units as available. Counting inbound units with no confirmed arrival date. Using a long velocity average that hides a recent change. Reviewing monthly on products whose lead time is measured in weeks. Treating overstock as harmless. It is slower-moving money and a growing storage bill. Related - Why Inventory Is the Second Pillar — why this outranks other optimisation work. - Inventory Forecasting and Purchase Orders — the planning half. - Catch a Stockout Early — the same check on the TrackIQ screen.

When You Are Going Out of Stock

The weekly check has told you a product will run out before the next shipment lands. The order is placed; it will not arrive in time. The instinct is to treat this as damage limitation. It is closer to the opposite: demand has outpaced supply, which is a position most sellers would pay for, and there is a way to convert it into unusually high profit while it lasts. Watch the walkthrough Prefer to watch? One hundred and twenty-six seconds on why running out means you are underpriced, the seven levers in the order that costs least, and the recovery step people neglect. The reframe You have a fixed number of units and more demand than units. That means you are currently underpricing — you are selling out at your current price, which is the market telling you the price could be higher. So the goal is not to sell the remaining stock as fast as possible. It is to sell exactly enough per day to reach the restock date, at the highest margin those units will bear. Step 1: work out the rate you need Simple arithmetic, and it sets everything else. Example. You planned for 200 units a day and are selling 275. At that rate you run out in 30 days. The next shipment lands in 30–35 days. Target: 175–200 a day. That stretches your cover to match the arrival window, with a margin for slippage. Always target slightly below the strict requirement. Shipments slip; demand spikes. Running out three days early undoes the whole exercise. Days of cover, and how fast it is falling Step 2: slow sales down, in order Work down this list, cheapest lever first. Stop when you hit your target rate. 1. Remove promotions and coupons. Free to reverse, immediate effect on conversion. 2. Pause outside traffic. Any external campaigns driving to the listing. 3. Cut advertising spend by half. Halve the product's target TACoS — see Setting Target TACoS per Product. 4. Remove Lightning Deals and Best Deals. 5. Pause advertising entirely. Set every campaign budget on the product to a nominal amount. 6. Raise the price. 7. Raise it again. The order matters. The first five are reversible with no lasting effect. Price changes are more visible to shoppers and to competitors, so they come last — but they are also where the profit is. Step 3: if it still sells too fast, keep raising the price Once advertising is off and promotions are gone, price is the only lever left. And this is the good outcome, not the bad one. If demand holds at a higher price, you are earning materially more on every unit for the rest of the window. Whether sales slow or not, you win: either you reach your restock date, or you make unusually high margin doing it. Keep going until the rate lands where you need it. Step 4: when stock lands, reverse everything — deliberately Undo the changes in reverse order. Price back first, then advertising, then deals, then coupons. Then push hard to return to your previous sales level. This is the step people neglect, and it costs them. Rank decays while a product is constrained. Coming back at your old price with your old budget will recover it slowly. Coming back with a coupon and elevated advertising recovers it much faster, because you need velocity to reclaim the position, not just availability. Treat it as a small relaunch — see Running a Launch: Stabilise, Scale, Scale Back. What this is worth Applied consistently across a catalogue, this changes the shape of a year rather than saving one product. The pattern is well documented by sellers who do it: a brand running high-teens to low-twenties net margin can push a strong month into the high twenties by applying this on every constrained listing at once — while still growing revenue, because the constrained products were selling out anyway. The money comes from not leaving the price where it was when demand outran supply. When not to do this When the stockout is avoidable. If you can expedite freight or move units from a 3PL, do that instead. This process manages an unavoidable shortage; it is not a substitute for supply. On a product mid-relaunch. You are deliberately buying rank, and throttling velocity to protect stock undoes exactly what you are paying for. Better to pause the launch and restart it when supply is secure. On a brand-new product still establishing rank and reviews, where continuity matters more than margin. Common mistakes Doing nothing and simply running out. The most expensive option available. Raising price first. The reversible levers are free; use them before the visible one. Cutting to exactly the required rate, leaving no room for the shipment to slip. Forgetting to reverse it and quietly running an over-priced, under-advertised listing for weeks after stock arrived. Coming back passively and letting rank recover on its own. Related - Monitoring Inventory Against Sales — the check that gives you enough warning to do this. - Inventory Forecasting and Purchase Orders — making it rare. - Catch a Stockout Early — spotting it on the TrackIQ screen.

Inventory Forecasting and Purchase Orders

Monitoring catches stockouts coming. Planning stops them happening. The planning half is a monthly habit rather than a system: forecast six months out, count everything you own wherever it currently sits, and order the difference. Done properly, stockouts stop being events. Watch the walkthrough Prefer to watch? One hundred and twenty-seven seconds on the six-month pipeline, the monthly ordering rule, and why over-ordering and under-ordering are not symmetrical mistakes. Build a six-month sales forecast Forecast six months ahead, per product, and update it monthly. Do not aim for precision. Nobody can predict Amazon exactly, and a forecast that is roughly right and actually maintained beats a precise one built once and abandoned. Where the numbers come from: Last year's figures for the same month, which is the only thing that captures your seasonality. Recent months against the same months last year, which tells you your growth rate. Weight them by how you are trending. Growing strongly? Lean on recent data, because last year understates you. Flat? Lean on last year, because it captures the seasonal shape better than a short recent window. Adjust for peaks, and be optimistic about them. Q4 numbers will look implausible next to a quiet spring. Order against them anyway — the asymmetry below explains why. Each month, roll it forward: revise the coming months with what you now know, and add a new sixth month. Sales history, the basis of any forecast Count everything in the cycle Your stock is not what is at Amazon. It is everything you have paid for or committed to, wherever it is: At Amazon. In your warehouse or 3PL. In transit. In production. A healthy pipeline, for a seller manufacturing overseas, looks roughly like: | Stage | Cover | |---|---| | At Amazon | 2–3 months | | Warehouse / 3PL | ~1 month | | In transit | ~1 month | | In production | 1–2 months | That is around six months in the cycle at any time. Manufacturing domestically? Same principle, shorter numbers — scale to your actual lead time. The monthly order rule At the start of each month, compare units in the cycle against your six-month forecast, and order the gap. If you are half a month short of six months' cover, place a purchase order for half a month. Not more, not less. Ordering monthly is the point. It keeps each order small, keeps the pipeline continuously topped up, and means a demand change is absorbed within weeks rather than at the next big order. If your supplier's minimum order quantity makes monthly impractical, tell them you intend to order every month — that regularity is often worth a lower MOQ to them. Failing that, order every two months and size accordingly. Adjust by product tier Not every product deserves the same cover. Top sellers: 5–6 months. These fund the business, and a buffer here is not optional. Everything else: 4–5 months, adjusted for cash flow. It is reasonable to run second-tier products leaner when cash is tight; it is not reasonable to do that with your best seller. Add a buffer, and make it real Add 0.5 to 2 months on top of your planned order. Delays are normal rather than exceptional — production slips, freight is delayed, receiving takes longer than expected. If you planned five months, order six. For top sellers, the buffer is non-negotiable. And a larger buffer creates room to grow. A product that cannot go out of stock can be advertised harder, relaunched, and pushed into a deal without anyone doing anxious arithmetic first. Where a product is performing well on advertising and rank, size the buffer generously — you are buying the ability to press the advantage. Plan Q4 in June Start Q4 planning in June and have the orders placed by the end of July. Include January and February in the Q4 order. Otherwise you land in Q1 having sold through everything with nothing behind it. If you manufacture in China, order March–June stock during October and November, because Chinese New Year closes factories for weeks and the queue afterwards is long. Sellers who miss this spend Q1 out of stock. Getting this right is one of the clearest advantages available, precisely because so many competitors leave it late. The asymmetry that should shape every decision If you place a £1,000,000 purchase order, you can sell at most £1,000,000 of it. You cannot sell more than you ordered. That is the whole argument for optimism. Over-ordering costs you storage fees and tied-up cash — real, but recoverable. Under-ordering costs you the sales, the rank, the reviews and the competitor's growth at your expense — and none of that is recoverable. The two errors are not symmetrical. Lean toward ordering more. What to review it against Forecasting needs actual sales history by product and month, current velocity, and stock across the pipeline. In TrackIQ that history sits alongside current stock, so the year-on-year comparison and the current runway are the same view rather than two exports. See Understanding Business Reports for sales history and How to Monitor Your Stock Levels for the pipeline. Common mistakes Forecasting once and never updating it. Counting only Amazon stock and being surprised by a gap the pipeline was supposed to cover. No buffer, so every delay becomes a stockout. Treating all products identically, and starving the top seller to fund the tail. Leaving Q4 to September, and Chinese New Year to January. Ordering pessimistically and capping your own year. Related - Monitoring Inventory Against Sales — the weekly half. - Why Inventory Is the Second Pillar — what a stockout actually costs. - Prepare for Prime Day — stock planning around an event.

Reimbursements, COGS and FBA Capacity

Three operational jobs that sit outside the weekly stock check, and quietly cost money when neglected. The first is the one most sellers do not know about at all. Watch the walkthrough Prefer to watch? One hundred and fifteen seconds on the sourcing cost figure every reimbursement is calculated from, why capacity requests are future-only, and how the two jobs compound. Keep your sourcing cost accurate Why this changed Amazon reimburses lost and damaged inventory based on your landed cost, not your selling price. That alone would be manageable. The problem is what happens when Amazon does not know your landed cost: it estimates one — from internal data, third-party sources, and broad assumptions about products in your subcategory. Those estimates are frequently wrong, and when they are wrong they are usually low. Every reimbursement you receive is then calculated from a number that undervalues your product. This is not a one-off task. It is a recurring monthly job, and accounts that skip it are consistently underpaid without ever seeing a line item explaining why. Check what Amazon currently thinks Seller Central → Inventory reimbursement → Cost management. Bulk Update → Download sourcing cost. Open the file and read the Latest Approved Cost column. That figure — not your selling price, not your invoice — is what every future reimbursement will be based on. The arithmetic is direct. If a shipment goes missing and you are reimbursed for 100 units, you receive 100 × that number. Compare against reality Work out your true landed cost per unit: manufacturing + freight + duties, divided by units. If Amazon's figure is lower than yours, update it. You are currently being under-reimbursed on every claim. If Amazon's figure is higher than yours, leave it alone. Amazon will reimburse above your actual cost. There is no obligation to correct a number in your favour, and no benefit in doing so. Then repeat monthly, because costs move — freight rates change, order quantities change, suppliers reprice. A figure that was accurate in March may be materially wrong by September. Inventory by status across the pipeline Request FBA storage capacity before you need it When to do it Capacity limits are the constraint nobody plans for until a shipment is refused. Requests are for future months only — you cannot request additional space for the month you are in. Which means the request has to be made before the need is obvious. So request ahead of Prime Day and Q4, at the same time you place those orders. If you are ordering Q4 stock in July, request the capacity in July. How Seller Central → Inventory → FBA Dashboard → Capacity Monitor, then Go to Capacity Manager and create a request. Select the size tier the product falls into — standard, oversize and so on are managed separately, and capacity in one does not help you in another. Choose the month you need the space for, and enter the additional volume in cubic feet. That last number is the one people get wrong. It is volume, not units — work it out from the carton dimensions of what you intend to send, not from the unit count. Claim shipment discrepancies The routine job When a shipment is received, Amazon's count and yours will sometimes differ. Units go missing in transit or in receiving, and the difference is only ever resolved if you raise it. Reconcile every inbound shipment against what Amazon recorded as received. Any shortfall is a claim. Keep the documentation. Claims are decided on evidence, so the invoice, packing list, carton counts and carrier documents need to be findable months later, not reconstructed under time pressure. And note the connection to the first section: a discrepancy claim is reimbursed at your sourcing cost. If that figure is understated, every successful claim is still underpaid. The two jobs compound. Why these belong together None of these is urgent in any given week, which is exactly why they get skipped — and each one quietly reduces what you recover or what you can hold. Put them on a schedule: Monthly — check and update sourcing cost. Monthly, alongside purchase orders — check capacity against what you are about to send. Per shipment — reconcile received against sent, and claim the difference. Where the numbers come from Updating sourcing cost requires knowing your true landed cost per unit, and reconciling shipments requires knowing what you actually sent against what arrived. TrackIQ holds the inventory side of that — units by status across your pipeline, and the sales history that tells you which products a reimbursement error would matter most on. See How to Monitor Your Stock Levels and Understanding Business Reports. Common mistakes Never checking what Amazon thinks your cost is. The most expensive omission here, because it silently reduces every claim. Correcting an estimate that was in your favour. Updating sourcing cost once and treating it as done. Requesting capacity in the month you need it, which is not possible. Entering units instead of cubic feet. Not reconciling inbound shipments, so shortfalls are never claimed. Related - Why Inventory Is the Second Pillar — where these jobs sit. - Inventory Forecasting and Purchase Orders — when to request capacity. - Prepare for Prime Day — the event these deadlines are usually driven by.

What AWD Is and When It Saves Money

AWD is Amazon's own bulk storage. You send inventory to an Amazon warehouse instead of your own 3PL, and Amazon distributes it into FBA from there. The reason to care is narrow and worth stating plainly: the end product is identical either way — units sitting in FBA. So the only question that matters is which route gets them there most cheaply. Watch the walkthrough Prefer to watch? One hundred and thirty-seven seconds on the two fees AWD avoids, the four it charges, and how much you save depending on your current route. Where the savings actually come from Two fees do most of the work. The inventory placement fee, charged when you send stock to too few fulfilment centres and Amazon redistributes it. The low inventory level fee, charged when your stock runs persistently thin relative to demand. AWD can avoid both. And because it also replaces a 3PL for everything except direct-to-consumer shipping, the saving often stacks. How much you save depends on where you are now | Current route | What AWD can save | |---|---| | China → 3PL → FBA | placement fees plus the 3PL-to-FBA leg | | China → FBA | placement fees | | Already on AWD | optimising carton size and units per case | The first row is where the money is. If you are paying a third party to store and then paying again to move stock into FBA, AWD collapses two costs into one. The third row is worth noting — sellers already using AWD frequently leave meaningful savings in how their cases are packed. See Working Out Whether AWD Is Cheaper. What AWD charges Four fees, and you need all four to compare honestly: Inbound processing — per case, to take stock into AWD. Outbound processing — per case, to send it out. Transportation — moving stock from AWD into FBA. Storage — monthly, on cubic feet occupied. Compare that total against shipping directly to FBA, including the placement fees you would pay on that route. It is not obvious in either direction without the arithmetic. Stock across the pipeline What AWD will not take Availability is still restricted, and the restrictions have tightened. Not accepted: dangerous goods and hazmat, with an exception for small and medium lithium batteries. Meltable and heat-sensitive products. Heavy or bulky items over 49 lbs or longer than 25 inches on any side. Now sortable items only. Since 31 July 2026, AWD accepts only items smaller than 18" × 14" × 8" and under 20 lbs. Anything larger goes directly into the fulfilment network through the standard send-to-Amazon workflow. That change removed a lot of products from AWD, so check your dimensions before planning around it. Expiry-dated products are now allowed, including health supplements, provided there are at least 185 days until expiration when the shipment is created. Where it fits with everything else Inventory planning. AWD is a stage in the pipeline, not a replacement for forecasting. It sits between production and FBA in the same cover calculation — see Inventory Forecasting and Purchase Orders. Low inventory fees. Avoiding these is a stated benefit, but the underlying cause is holding too little stock relative to demand — see Reducing FBA and Referral Fees. Large B2B orders. AWD works at case level rather than unit level, which suits customers buying full cases — see B2B Fee Discounts, Certifications and Fulfilment. Do the sums before committing AWD is not automatically cheaper. It is cheaper for some products on some routes with some packing configurations, and more expensive otherwise. The next article is the arithmetic — see Working Out Whether AWD Is Cheaper, and AWD Rate Tiers and Eligibility for the discounts that change the answer substantially. Related - AWD Rate Tiers and Eligibility — the three discount tiers and what they require. - Working Out Whether AWD Is Cheaper — the comparison, step by step. - Inventory Forecasting and Purchase Orders — where AWD sits in the pipeline.

AWD Rate Tiers and Eligibility

AWD's list prices are not what most sellers pay. Three discount tiers can take a meaningful bite out of storage and transport — and they stack differently depending on how you ship and how you replenish. Understanding them changes the answer to "is AWD cheaper", sometimes decisively. Watch the walkthrough Prefer to watch? One hundred and twenty-six seconds on the three discount tiers and what each requires, and how to steer auto-replenishment instead of overriding it. The Integrated Rate: 25% off The easiest discount to earn. Ship into AWD using Amazon Global Logistics or the Partnered Carrier Program, and you qualify for rates 25% below standard. The requirement is lighter than it sounds: at least one shipment into AWD via AGL or PCP in the previous four calendar months. Eligibility is assessed on the 25th of each month, applying to the following month's charges. Which means roughly three qualifying shipments a year keeps you permanently on the lower rate. If you are already sending several shipments annually, routing a few through Amazon's own carriers is close to free money. Smart Storage: 10% off storage Rewards letting AWD replenish FBA automatically rather than doing it by hand. The requirement: at least 70% of the SKU's units replenished to FBA over the last 90 days must have been automatic. Put differently, manual replenishment must account for under 30% of units. Also required: more than 70 days of historical demand, measured across combined FBA and AWD inventory — the same test used for the low inventory level fee, assessed on the 25th. How to keep it without giving up control The obvious problem: you sometimes need to push stock manually — before a relaunch, ahead of peak season, when stock runs short. The technique is to stop replenishing manually and start steering the automation instead. Adjust the minimum and maximum guardrails on auto-replenishment rather than creating manual shipments. Worked example. A product normally sells about 100 units a month, with guardrails around 80 and 200. About a month before Prime Day or Black Friday, raise them to 200 and 400. AWD then auto-replenishes to the higher level on its own. Check over the following days that a shipment was actually created. The guardrail change is the instruction; the shipment is the evidence it worked. This gets you the stock you wanted while keeping the replenishment automatic, so the 70% threshold holds and the discount survives. Keep manual replenishment for the genuine exceptions — a relaunch, an unexpected spike, a stock emergency. Amazon Managed Rate: 20% off storage, 10% off transport The largest discount, and it requires everything above plus one more thing. An Amazon partner carrier — AGL, SEND or PCP. The same 70% auto-replenishment threshold. The same 70+ days of historical demand. So the tiers compound. Ship through Amazon's carriers, let auto-replenishment do the work, and keep a reasonable demand history, and you are on the best available rates on both storage and transport. Days of cover across the pipeline The fee structure changed in 2025 Worth knowing if you are working from older figures: processing fees were split into separate inbound and outbound charges in April 2025. Comparisons built before that will understate your costs. What to check before planning around AWD Is your product eligible at all? Since 31 July 2026, AWD is sortable items only — under 18" × 14" × 8" and under 20 lbs. Plus the standing exclusions on hazmat, meltables and heavy or bulky goods. Do you have 70+ days of demand history? New products do not, which means the storage discounts are unavailable until they do. Can you route shipments through AGL, SEND or PCP? If not, the Integrated and Managed rates are both closed to you, and your AWD comparison should use standard rates. Are you replenishing manually out of habit? If most of your replenishment is manual and could reasonably be automatic, you are paying more for storage than you need to. How this changes the decision Run your AWD comparison at the rate you will actually pay, not the list rate. A product that looks marginally more expensive on AWD at standard rates can be clearly cheaper on the Managed Rate. Getting the tier wrong in the calculation gets the decision wrong. See Working Out Whether AWD Is Cheaper. Related - What AWD Is and When It Saves Money — the fees and the eligibility rules. - Working Out Whether AWD Is Cheaper — the comparison these rates feed. - Inventory Forecasting and Purchase Orders — the demand history these thresholds assume.

Working Out Whether AWD Is Cheaper

AWD is cheaper for some products and more expensive for others, and you cannot tell which from the fee schedule alone. The comparison needs real shipping quotes on both routes. It takes an afternoon per product group, and it settles a decision worth thousands a year. Watch the walkthrough Prefer to watch? One hundred and twenty-three seconds on the five-step comparison, why units per case can decide it alone, and why the invoices are worth reconciling. The five steps 1. Get your real product data Dimensions and weight, from Manage All Inventory. Use Amazon's recorded figures, not your own measurements — Amazon charges on what it recorded, and the two are not always the same. If those figures look wrong, that is a separate saving worth claiming — see Reducing FBA and Referral Fees. 2. Price the direct-to-FBA route Create a dummy shipment into FBA, choose the cheapest option, and take the fulfilment centre address to your freight agent for a quote. A dummy shipment is the only way to get the real destination, because Amazon assigns the centre and that assignment drives the freight cost. 3. Price the AWD route Create a dummy shipment into AWD, same unit count, and get that shipping cost. Get both quotes for the same volume. Different unit counts produce different rates, and comparing across them is the most common way this exercise goes wrong. 4. Put both into the comparison You are comparing total landed cost into FBA, both routes, including: Freight to the destination. Inbound processing (AWD only). Storage for the period you expect to hold. Outbound processing and transportation into FBA (AWD only). Placement fees you would pay going direct. Use the rate tier you will actually qualify for — standard, Integrated, or Managed. See AWD Rate Tiers and Eligibility. 5. Decide per product group The answer varies by product, so do not generalise from one. Small, dense, fast-moving products behave very differently from large, slow ones. The lever most people miss: units per case Savings rise substantially with how many units you fit in a case, because inbound and outbound processing are charged per case, not per unit. Twice the units per case means half the processing cost per unit. So work out the maximum that will physically fit, then have that conversation with your supplier: 1. Ask whether the carton size can change at all. Often it can; nobody has ever asked. 2. If yes, ask how many units fit in the largest workable carton. 3. Push toward the theoretical maximum, not just an improvement. 4. Then compare old units-per-box against new and price the difference. This frequently changes the decision on its own. A product that is marginal on AWD at 24 units per case can be clearly cheaper at 48 — without renegotiating a single rate. And it helps on the direct route too, since better-packed cartons reduce freight cost per unit regardless of where they go. Units and cost, per product What tips the answer toward AWD You currently use a 3PL, so AWD collapses storage and the transfer leg into one cost. You pay placement fees regularly. You can hit the Managed or Integrated rate. Your product packs densely into large cartons. You hold several months of cover, so storage efficiency matters. What tips it against You ship direct to FBA already and rarely pay placement fees. Your product is bulky or light, so cases hold few units and processing dominates. You hold little stock, so storage savings are small. You cannot use Amazon's carriers, closing off both discount tiers. Your product is not eligible — and since 31 July 2026 that excludes anything non-sortable. Then keep checking the bills AWD is a newer service and the fee reconciliation is worth doing rather than assuming. Amazon provides reports covering inventory and fees charged. Reconcile them — check the cases you sent against the cases processed, and the storage charged against the volume you actually occupied. Discrepancies happen, and they are claimable. This is the same discipline as reconciling inbound FBA shipments — see Reimbursements, COGS and FBA Capacity. Common mistakes Comparing fee schedules instead of getting real freight quotes. Quoting different unit volumes on the two routes. Using list rates when you would qualify for a discount tier. Ignoring units per case, which is often the largest single lever. Deciding once for the whole catalogue rather than per product group. Never reconciling the invoices. Related - What AWD Is and When It Saves Money — the fees involved. - AWD Rate Tiers and Eligibility — the rate to run the numbers at. - Negotiating Lower COGS — the same supplier conversation, wider scope.

AWD Shipping and Replenishment

Once you have decided AWD is worth using, the operational side is straightforward — it works almost exactly like inbounding to FBA. There are two places it differs, and both matter. Watch the walkthrough Prefer to watch? One hundred and sixty-three seconds on checking regional restrictions before you build a shipment, the partnered-carrier discount, and replenishing FBA in time. Where it lives Menu → Inventory → AWD (Warehousing and Distribution). Three things happen here: view inventory, track shipments, create shipments. Creating a shipment into AWD Identical to a normal FBA inbound. Enter the boxes you are ready to send, create labels, ship. The one real difference is destination. Regional destinations, and their restrictions AWD assigns you to a region — East Coast, West Coast, South Central or Southeast — and each has its own acceptance rules, which change as the network expands. In practice this has meant pallets to some regions and parcels to others, with some regions closed to new shipments entirely. So check the current restriction before you build the shipment, not after. Discovering your region will not take parcels when the boxes are already labelled is an expensive afternoon. Choose Amazon-assigned unless you have a reason not to, and let Amazon tell you where the inventory should go. And know your freight position. If you are using Amazon's own freight the cost is unaffected by which region you are sent to. If you are shipping independently, the destination changes what you pay — so a region assignment is a cost decision, not just a logistics one. The fee lever: use an Amazon-partnered carrier Storage and processing fees are lower when the inbound is Amazon-managed. Choose a partnered carrier at shipment creation. It is a one-click difference at setup and a permanent difference on every unit stored — see Is AWD Cheaper?. Viewing what is there Inventory → Warehousing and Distribution shows, per product: Units at AWD. Units in transit. Units with errors. Watch the errors column. It is the only place some receiving problems surface before they become missing inventory — see AWD Reimbursements. Moving inventory from AWD onward Move from AWD, then select the products. Two destinations: FBA — the normal replenishment path. A retail or distribution partner — add the channel and Amazon ships to them. You get a cost before you confirm, which is the number to check against your own freight alternative rather than accepting by default. Track outbound shipments in the same place. They also appear on the normal FBA Shipments tab once they are moving, so you are not tracking them in two systems. Inventory position across AWD and FBA Once inventory sits in two places, "how many units do I have" stops being one number — see Monitoring Inventory Against Sales. Cover calculations that ignore AWD stock will tell you to reorder when you already hold three months of it. How replenishment changes your planning AWD is a buffer, not a warehouse you forget about. Your FBA cover calculation now has a second lead time in it — the AWD-to-FBA replenishment, which is days rather than the weeks an overseas shipment takes. Which is the whole point. You can hold less in FBA, avoid FBA storage fees on slow-moving stock, and still replenish quickly — see Inventory Forecasting and Purchase Orders. But it only works if you actually trigger the replenishment in time. Stock sitting in AWD while the FBA listing goes out of stock is the worst of both arrangements — see Going Out of Stock. Common mistakes Building a shipment before checking the region's current restrictions. Not choosing an Amazon-partnered carrier, and paying higher storage and processing fees permanently. Ignoring the errors column in the inventory view. Counting only FBA units when calculating cover. Treating AWD as long-term storage rather than as a replenishment buffer. Leaving replenishment too late, and going out of stock with inventory 200 miles away. Related - What AWD Is — whether to use it at all. - Is AWD Cheaper? — the cost comparison this operational work rests on. - AWD Reimbursements — reconciling what you sent against what arrived.

AWD Reimbursements

Inventory moving through AWD crosses two hand-offs — your factory or 3PL into AWD, and AWD into FBA — and units go missing at both. Sellers have been losing meaningful quantities here, and Amazon does not reimburse what you do not claim for. So this is a monthly reconciliation, not an occasional check. Watch the walkthrough Prefer to watch? One hundred and seventy-one seconds on reconciling inbound and outbound shipments, filing a claim that gets paid, and working around the 30-day report limit. Step 1: reconcile the inbound Seller Central → AWD → Reports → Inbound shipment report. Select Last 30 days and download. Then three moves: 1. Filter shipment_status to Closed. Closed means Amazon has finished processing and will make no further updates — anything else is still in flight and not yet a discrepancy. 2. Compare expected_unit_count against received_unit_count. 3. Where received is lower than expected, Amazon has lost or misplaced those units. Do not chase shipments that are still Receiving. The numbers move until the shipment closes, and a claim on an open shipment wastes a case. Step 2: reconcile the outbound Seller Central → FBA Inventory → dropdown → Download shipment details (AWD to FBA). Same logic: Filter to Closed shipments. Compare Confirmed quantity against Received quantity. Any gap is missing inventory. There is also an AWD Inventory Report available from the same dropdown, which gives a snapshot of units at AWD and units outbound to FBA. Useful for a position check; the shipment detail report is what you claim from. Step 3: file the claim Open a case with Seller Support and give them everything at once — a case missing an identifier fails slowly. See Escalating an Amazon Case. For an inbound discrepancy: Shipment ID, ASIN, Amazon reference ID, FNSKU, and the number of missing units. For an outbound discrepancy: Shipment ID, SKU, and the number of missing units. Attach the product invoice as supporting evidence in both cases. Claims are routinely rejected for its absence. A template that works Hi team, There is a discrepancy between units sent and units received for ASIN xx (SKU xxx) from Shipment ID yy. Shipment ID: __ Amazon reference ID: __ ASIN: __ FNSKU: __ Number of missing units: __ Please research the missing units from this shipment. If the units have been misplaced or lost, we are requesting a reimbursement. I have attached the invoice for reference. Thank you. Note what it does: states the discrepancy, gives every identifier, and asks for a specific action. The 30-day trap The AWD replenishment report only covers the last 30 days. Which means a discrepancy you do not catch inside a month disappears from the report entirely. Two consequences: Run this monthly, without fail. A quarterly reconciliation is a reconciliation that misses two months of claims. Audit older shipments manually on the AWD-to-FBA page, shipment by shipment, until Amazon's reporting improves. Tedious, but it is the only route to anything beyond the 30-day window. This is a good process to delegate — it is well-defined, repetitive, and its output is a case you review before filing. See Onboarding a VA. While you are there: check the fees Three AWD fees, three reports, and each is worth confirming rather than assuming: Storage — charged on the 6th of each month, on average volume stored across the month. Read total_charged_amount on the Consolidated Report tab of the Monthly Storage Fee Report, broken out by warehouse. Processing — charged per box when a replenishment is delivered. Sum calculated_fee_amount in the Monthly Processing Fee report. Transportation — charged on cubic feet delivered from AWD to FBA. Same column, in the Monthly Transportation Fee Report. Check these against the rates you expected — see AWD Rate Tiers and Is AWD Cheaper?. If the rates have drifted, the AWD-versus-FBA calculation that justified using it may no longer hold. Fees and units, per product Fees that drift and units that quietly go missing both look like normal cost until something compares them against what they should be — see Reimbursements, COGS and Capacity. Common mistakes Claiming on shipments that are not yet Closed. Filing without the invoice, and getting rejected. Running the reconciliation quarterly, so most discrepancies age out of the report. Reconciling inbound only, and ignoring the AWD-to-FBA leg where losses also occur. Never checking the fee reports, and assuming the rates you were quoted are the rates you pay. Treating this as optional. Unclaimed units are a permanent write-off. Related - AWD Shipping and Replenishment — the operations this reconciles. - Reimbursements, COGS and Capacity — the wider reimbursement discipline. - Escalating an Amazon Case — what to do when a claim stalls.

Requesting FBA Capacity

Your FBA storage limit is what caps how much stock you can hold — and going into Prime Day or Q4 with a limit that will not fit your plan is a self-inflicted stockout. Capacity Manager is how you bid for more, and it is a bid, not a request: you name a maximum price per cubic foot and Amazon allocates space accordingly. It is also a well-defined, delegable process — see Onboarding a VA. Watch the walkthrough Prefer to watch? One hundred and seventy seconds on converting forecast units into cubic feet, bidding against recent winning rates, and resubmitting after a denial. Where it is Inventory → FBA Dashboard → Capacity Monitor (at the bottom of the screen). Choose your product size tier — standard-size, oversize and so on — then Go to Capacity Manager, and Create a new request. Requests are per size tier. If you sell both standard and oversize products, you need a request for each. Step 1: pick the month You cannot request for the current month — only upcoming ones. Which is the single most important operational fact here. Realising in late October that you need November space is already too late for October, and the shipment has to arrive before the space matters. Work backwards from your inbound lead time, not from the month you need the stock. Step 2: work out how much space you actually need Translate forecast units into cubic feet. Case pack cubic feet = (L × W × H in inches) ÷ 1,728. Total cubic feet = (forecast units ÷ units per case) × case pack cubic feet. Worked example 5,000 units of a lotion forecast for Q4. 12 units per case, case dimensions 16" × 12" × 10". Case volume: (16 × 12 × 10) ÷ 1,728 = 1.11 cu ft Cases needed: 5,000 ÷ 12 = 417 Total: 417 × 1.11 = 463 cu ft Check your size tiers. Bulky products consume space far faster than the unit count suggests, and a tier misclassification throws the whole calculation. Then request 20–30% more than the calculation — but not dramatically more. Amazon charges reservation fees on space you leave unused, though performance credits can offset this if your sales volume is strong. The forecast is the input to all of this — see Inventory Forecasting and Purchase Orders. Forecast units, translated into space you have to book A capacity request is only as good as the forecast under it, and the forecast has to be per product, seasonal, and current — see Monitoring Inventory Against Sales. Requesting against last year's numbers is how sellers end up paying reservation fees on space they never fill. Step 3: set your bid The Maximum Reservation Fee is the most you will pay per cubic foot if you actually use the space. Do not take Amazon's recommendation at face value. Use Download Reservation Fees History to see what sellers recently paid. If Amazon recommends $2.00 but recent winning bids were $4.00–6.00, bid competitively — a low bid is simply a request that quietly fails. The panel shows your total maximum exposure — 400 cu ft × $2.00 = $800. Check that number before submitting, because it is what you are committing to. Step 4: set an expiration date If you need a decision by a deadline, set one. Give yourself 1–2 weeks of margin before the shipment has to move. An approval that lands the day your freight was due to leave is not an approval you can use. Step 5: review and submit Check the requested volume and the total maximum reservation fee against your actual plan, then submit. If it is denied, edit and resubmit — raise the bid, or adjust the volume. A denial is a negotiating position, not a final answer. If Amazon shows you the "I need help" screen instead Some accounts see a guided version rather than the manual form. Choose "I need help" and Amazon estimates the capacity and suggests a bid from your forecast sales. But treat those numbers as a starting point. They tend to be conservative in a specific direction — a higher bid and a lower volume than you probably need. Edit both before submitting, using your own cubic-feet calculation. Common mistakes Requesting in the month you need the space, which is not possible. Forecasting units without converting to cubic feet. Bidding at Amazon's recommendation without checking recent winning bids. Requesting far more than you need, and paying reservation fees on empty space. Requesting for one size tier when your catalogue spans two. Treating a denial as final rather than resubmitting with a higher bid. Leaving no margin on the expiration date, so approval arrives too late to ship against. Related - Inventory Forecasting and Purchase Orders — the forecast this request is built on. - Prepare for Prime Day: Budget Rules and Inventory Cover — the event this usually exists for. - Reimbursements, COGS and Capacity — the wider capacity and cost picture.

Keeping Your Sourcing Cost Current

Since 2025, Amazon reimburses lost and damaged inventory at your landed cost, not your selling price. But Amazon does not use your actual cost. It estimates one, using internal data, third-party sources and broad assumptions about products in your subcategory. Those estimates are frequently wrong, and usually wrong downwards — which means every reimbursement you receive is quietly short. So keeping your sourcing cost current is a monthly process, not a one-off. Watch the walkthrough Prefer to watch? One hundred and sixty seconds on reading Amazon's approved cost, correcting only in the direction that pays you, and the invoice detail that decides whether it is approved. Step 1: find out what Amazon thinks your cost is Seller Central → Inventory Reimbursement → Cost Management. Bulk Update → Download sourcing cost. Open the CSV and read column D, Latest Approved Cost, for each listing. That number is the basis of every reimbursement you will receive from now on. A shipment where 100 units go missing pays out at 100 × that figure — not your invoice, not your selling price. Step 2: compare it against reality Put Amazon's figures alongside your actual landed costs, per SKU. Landed cost means everything to the warehouse door — unit cost, freight, duty — see Where Your Margin Actually Goes. Then read the difference: Amazon's figure is lower than your cost → you are being underpaid. Fix it. Amazon's figure is higher than your cost → Amazon will pay you more than the units cost you. Do nothing at all. Correcting this downwards costs you money for no benefit. That asymmetry is the whole reason this is worth doing carefully rather than in bulk. You are correcting one direction only. Cost and units, per product This comparison needs a current, per-SKU landed cost, which is the same figure your margin analysis runs on — see Understanding Business Reports. Stale COGS produces both an understated reimbursement and an overstated margin, from the same error. Step 3: submit the correction Where column D is below your actual cost, enter your real COGS in column H. Save, upload with the Upload file button, then Submit for review. Amazon takes a few days. Track status by SKU on the same Cost Management page. Step 4: be ready to prove it Amazon frequently asks for evidence — for example, when its estimate says $27.39 and you have claimed $32.49. Click Update and upload an accepted document: a commercial invoice or wholesale invoice. It must show the product name and the Amazon SKU. If your supplier's invoices do not, ask them to reissue with both. This is the single most common reason a correction stalls, and it is fixable once, permanently, by telling your supplier what the invoice needs to contain. Once accepted, the status changes to Approved and future reimbursements use your figure. Make it a standing monthly process Costs change. Freight rates move, suppliers re-quote, duty changes — and Amazon's estimate does not follow any of it. So this belongs on the monthly calendar — see The Brand Manager's Calendar. And it is an ideal delegation: repetitive, well-defined, with a clear correct answer. Assign one owner. Give them this process and confirm they understand it. Fix the exact date each month. Walk them through the first run, so the first submission is right. See Onboarding a VA. The other half: actually claim Correct costs only pay off if you are filing the claims. Track shipments weekly and file for discrepancies — see Reimbursements, COGS and Capacity and AWD Reimbursements. At real shipment volume the combined amount is significant, and both halves have to be running: the claim, and the cost basis it pays out on. Common mistakes Assuming Amazon uses your cost. It uses an estimate. Correcting Amazon's figure downwards when it is higher than your actual cost. Submitting invoices without the Amazon SKU on them, and stalling the review. Doing it once, then letting it go stale as costs move. Updating costs but never filing claims, or filing claims on a stale cost basis. Keeping it in your own head rather than assigning it a named owner and a date. Related - Reimbursements, COGS and Capacity — the claims this cost basis pays out on. - AWD Reimbursements — the same discipline for AWD shipments. - Where Your Margin Actually Goes — where your real landed cost comes from.

Creating a Shipping Plan

A shipping plan tells Amazon what you are sending, how much, and how it is packed. Nothing reaches FBA without one, and most receiving problems are created here rather than in transit. The errors are all preventable, and they all cost the same thing: delay, at the exact moment you were trying to avoid a stockout. Watch the walkthrough Prefer to watch? One hundred and eighty-four seconds on the settings that decide your freight cost, choosing a mode against your remaining cover, and the details that cause check-in failures. Creating it Inventory → Shipments (under Fulfilment by Amazon) → Send to Amazon. Step 1: shipping details Ship-from address — your supplier, your 3PL, or your own warehouse. This is not administrative. The origin influences which Amazon warehouses your stock is assigned to, and therefore what the freight costs. Marketplace destination — check it matches where you intend to sell. Then one checkbox that catches people out. Do not tick "I want to ship with Amazon or an Amazon-partnered carrier" unless you actually are. Using your own freight forwarder means leaving it unchecked. Step 2: products and packing Select the SKUs, then per SKU: Units per box. Box dimensions. Box weight. Use case packs — one box, multiple units of the same product. It is faster to check in and far less error-prone than mixed boxes. Then enter the number of boxes, and Amazon calculates the total units. Step 3: confirm — where the errors surface Amazon checks the shipment against your storage capacity limits. Over the limit and you cannot proceed — see Requesting FBA Capacity. It also flags hazmat requirements and missing compliance or certification. Fix everything shown before continuing. These do not resolve themselves later. Step 4: ship date, mode, and carrier Ship date is when you hand the goods to the carrier, not when they arrive. Small Parcel Delivery or pallets | | Small Parcel (SPD) | Pallets | |---|---|---| | Format | boxes only | palletised | | Best for | small shipments, a few cartons | bulk volume | | Cost per unit | higher | lower | | Check-in speed | fast — days | slow — up to two weeks | | Freight | parcel carrier, partnered or your own | LTL/FTL, usually a forwarder | Rule of thumb: a few boxes → SPD. Bulk inventory → pallets. And use SPD deliberately as a stockout tool. When a product is days from going dark, a small parcel shipment checked in within days can bridge the gap that a pallet shipment cannot — see Going Out of Stock. Amazon then assigns the fulfilment centres, and may split your shipment across several. That is normal, and it changes your cost. If you use an Amazon-partnered carrier, the shipping fees are charged to your Amazon account balance once you proceed. Review the estimated fees before confirming, not after. Step 5: labels Print box labels. Every box must carry the correct label for its SKU. Have your supplier apply them before shipping. It is cheaper and faster than having a 3PL do it on arrival. Missing or wrong labels is the most common cause of check-in problems, and it is entirely avoidable. Step 6: pallets, if applicable Enter the pallet count and print pallet labels. Palletise in the destination country, through your forwarder or 3PL. Palletising at origin wastes container space and raises your freight cost — you are shipping air. Step 7: tracking Add a tracking ID for parcel shipments, or a BOL / freight reference for pallets. Some marketplaces also want customs or import references once the goods land. Amazon needs this to receive the shipment properly, so accuracy matters more than it appears. Seven things that cost people money here Created shipments count against your capacity immediately. Cancel anything you are not going to send, or it is consuming space you may need. Shipments expire after about 90 days. Ship and deliver inside that window. Send exactly what you declared. Mismatched units, box counts, dimensions or weights cause delays, extra fees, or refusal. Labels are mandatory and must be correct. Monitor until check-in. A stuck shipment is only expensive if nobody notices. Expect split shipments, and factor the cost. Consider AWD for bulk, where it is often cheaper than sending everything straight to FBA — see Is AWD Cheaper? and AWD Shipping and Replenishment. Cover per product, and how long the inbound has Which mode you choose is really a question about how many days of cover you have left, and that has to be current per product — see Monitoring Inventory Against Sales. A pallet shipment against three weeks of cover is a stockout you have already scheduled. Common mistakes Ticking the partnered-carrier box while using your own forwarder. Creating shipments you do not send, and eating your own capacity. Sending different quantities from those declared. Labelling at the 3PL rather than at the supplier. Palletising at origin, and paying to ship empty space. Choosing pallets when cover is short, and going out of stock during check-in. Not tracking through to check-in, so a stuck shipment surfaces weeks late. Related - Requesting FBA Capacity — the limit this shipment is checked against. - AWD Shipping and Replenishment — the alternative route for bulk inventory. - Inventory Forecasting and Purchase Orders — deciding what to send and when.

Claiming for Shipment Discrepancies

Units go missing between your carrier and Amazon's shelves. Amazon reimburses for them — but only on shipments it has finished processing, and only when you ask. The process is fiddly rather than difficult, and the two places it fails are the documents and the follow-up. Automated reimbursement services exist and are worth considering if the volume justifies it. This is the manual process, which you need to understand either way. Watch the walkthrough Prefer to watch? One hundred and sixty-four seconds on finding eligible closed shipments, the invoice and BOL requirements, and the two-day response window that closes cases. Step 1: find the eligible shipments Inventory → Shipments. Filter Status to Closed and apply. Closed matters: Amazon only reimburses once it has completed processing. An open shipment's numbers are still moving. Then look for shipments where Units located is less than Units expected — 1,008 shipped, 989 located, for example. That 19-unit gap is your claim. Step 2: open the reconciliation Track Shipment → Contents. Review the discrepancy. If it is negative and the status is Action Required, proceed. Then choose Research Missing Units from the dropdown, for each SKU with a gap. Step 3: the documents — where most claims fail Two documents, both required. A commercial invoice, under Proof of Inventory Ownership It must show: Invoice number. Purchase date — dated at least a month before the shipment creation date. A purchase date after the shipment was created makes no sense and gets the claim rejected. Manufacturer name and address. Buyer name and address. Ship-to location. Every product name on the shipment summary, with expected quantities. Incoterms. The Bill of Lading, under Proof of Delivery Then confirm both appear under Attached Documents before submitting. Uploads silently fail more often than they should. Step 4: the message Keep it short and factual: Dear Amazon Seller Support, There is a discrepancy between units sent and received for shipment ID [ ]. Please investigate the missing units from this shipment. Thanks! Regards, Team [Brand] Preview Your Request, then Submit. A pop-up confirms the case and gives you a Case ID. Step 5: record it Keep a tracking sheet. At minimum: shipment ID, SKU, units missing, Case ID, date opened, status, result, remarks. Record the Case ID and set the status to Case Opened immediately. Without it you cannot follow up, and follow-up is the part that gets claims paid — see Escalating an Amazon Case. Step 6: the deadline that closes cases This is the detail that costs people money. Amazon may ask for additional documents. Provide them within two days. Failure to respond within five days closes the case automatically. And if you have heard nothing two days after submitting, follow up — Help → Manage Support Cases, find the case by ID, and reply on the thread. A closed-for-non-response case is not reopened. You start again, if the shipment is even still eligible. Which is why this needs an owner and a routine, not good intentions — see Onboarding a VA. When it is approved, record the reimbursement ID and close the row. Where Amazon fails to investigate properly, note that explicitly — the pattern across cases is itself useful. Checking what you have actually been paid Seller Central → Inventory Reimbursement → Overview. It summarises reimbursements for inventory defects — lost, misplaced and similar — over a date range, and can be filtered to a single SKU. Reconcile that against your tracking sheet. Cases marked reimbursed that never produced a payment do happen. Units and cost per product, which is what a claim pays out on And remember what the payout is calculated on: Amazon's estimate of your landed cost, not your invoice — see Keeping Your Sourcing Cost Current. Filing perfect claims against a stale cost basis means being paid promptly and incorrectly. Make it weekly Shipments close continuously, and eligibility does not last forever. A weekly pass over closed shipments is enough to catch everything, and at real shipment volume the annual total is significant. Run the same discipline on AWD shipments, which have their own reports and their own 30-day trap — see AWD Reimbursements. Common mistakes Claiming on shipments that are not Closed. An invoice dated after the shipment was created. Missing the Bill of Lading, or not confirming both uploads attached. Not recording the Case ID, and losing the ability to follow up. Missing the two-day document window, and having the case closed automatically. Never reconciling approved cases against actual payments. Filing against a stale COGS figure, and being underpaid on every successful claim. Related - Reimbursements, COGS and FBA Capacity — where this sits in the wider picture. - Keeping Your Sourcing Cost Current — the cost basis these claims pay out on. - Creating a Shipping Plan — the shipment these discrepancies come from.

Inventory Placement: Split or Pay

Amazon gives you a choice on every inbound shipment: send everything to one location and pay an inventory placement fee, or split it across several fulfilment centres and pay no fee. Most sellers pick one and never revisit it. But which is cheaper is a per-product calculation, and it flips depending on how your product packs. Watch the walkthrough Prefer to watch? One hundred and sixty-three seconds on why packing efficiency decides the answer, costing both options as dummy shipments, and the costs that never appear in a freight quote. The trade Pay the placement fee — one destination, one freight movement, a fee per unit. Split the shipment — no fee, but you are now paying freight to several destinations instead of one. The split option is only cheaper when the extra freight costs less than the fee. That depends almost entirely on how efficiently your product fills a box. Why packing efficiency decides it A product that fits 48 units to a box where the box could hold 182 is shipping mostly air. And you cannot always fix that — case pack configurations are frequently fixed by your supplier and not negotiable. When a product packs badly, every additional destination multiplies an already inefficient shipment, and splitting across five locations gets expensive quickly. When a product packs densely, the extra destinations cost comparatively little, and avoiding the fee wins. So the answer is genuinely per product, and worth recalculating when case packs or rates change. How to work it out Build the same shipment twice as a dummy, and compare. 1. Create a dummy inbound to FBA with the split — Amazon assigns the destinations and quotes the freight. That gives you the multi-destination cost. 2. Create the equivalent single-destination shipment and read the placement fee. 3. Compare, and divide by units to get the saving per unit. A worked example A product with a warehouse in California, 48 units per box. Shipping to AWD in California: quotes ranged $430–620, so take a midpoint of about $500. The same shipment inbound to FBA across five locations: $1,543. Splitting to five FBA locations was substantially more expensive than a single AWD movement — so for this product, the consolidated route won clearly. That gap is a packing problem showing up as a freight problem. Units and cost, per product This calculation is per product, per shipment, and it changes as rates and case packs change — see Where Your Margin Actually Goes. A decision made once against last year's rates is not a decision that still holds. The third option AWD often beats both. One inbound movement to a single AWD facility, then Amazon distributes to FBA — so you avoid the placement fee and the multi-destination freight, and you gain a replenishment buffer. Run the comparison three ways, not two: Split FBA shipment, no fee. Consolidated FBA shipment, plus fee. Inbound to AWD, plus AWD storage, processing and transportation fees. See Is AWD Cheaper? and AWD Shipping and Replenishment. What else the split costs you Freight is not the only difference. Check-in time. More destinations means more shipments to track, and any one of them can stall — see Creating a Shipping Plan. Reconciliation effort. Each destination is a separate shipment to check for discrepancies — see Claiming for Shipment Discrepancies. Stock fragmentation. Units spread thinly across regions can leave you technically in stock but slow to serve. None of these appear in the freight quote, and on a marginal calculation they should tip you toward consolidation. Common mistakes Choosing once and applying it to every product regardless of how they pack. Comparing the fee against freight to one destination rather than to all the destinations Amazon assigns. Ignoring case pack efficiency, which is the variable that actually decides it. Leaving AWD out of the comparison, where it frequently wins. Counting only freight, and ignoring the tracking and reconciliation load of five shipments. Not recalculating when rates move or a supplier changes the case pack. Related - Is AWD Cheaper? — the third route, and how to cost it. - Creating a Shipping Plan — where the split is assigned. - Where Your Margin Actually Goes — what a per-unit freight saving is worth.