Inventory and Supply
The chain, not the spreadsheet.
Selling rate decides what you are short of. Being short decides what you order. What you order becomes a lot, and the lot becomes the shipment Amazon is expecting. Every link is usually somewhere else.
- Unit of planning
- The Sales SKU, with its own buffer and rate
- Ends at
- An FBA or AWD shipment with an arrival date
Inventory and Supply
You sell first. Then you make more.
Nobody manufactures a year of stock and hopes. You sell, you watch the rate, and you place the next run in time for it to land before you run out. That is a chain of decisions, and every link of it is somewhere else: a spreadsheet, an email to the factory, a shipment ID you paste into Seller Central.
- Real demandSelling rate per week, read across several windows
- ProjectionStock on hand, inbound, FBA and AWD, week by week
- Shipment planWhat is short, against the cover you want to hold
- Purchase orderThe approved plan, with your manufacturer’s minimums checked
- AllocationWhat the factory made, assigned to the plans that need it
- Tracked lotMapped to the real shipment, with an arrival date
Material planningThe purchase order proposes the components it needs, projected week by week against what is on hand and on order, per supplier. For sellers who buy the parts, not just the box.
The week your stock runs out, before it does.
Every Sales SKU gets a projected end-of-week stock line: what you hold in FBA and AWD, what is inbound and when it lands, against the rate you are actually selling at.
| Sales SKU | Coverage | Sales / week | Plan & arrivals | W36 | W37 | W38 | W39 | W40 | W41 | W42 | W43 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| HRB-1P | 8w buffer | 66 | Planned +1,080 | 365 | 299 | 342 | 276 | inbound1,290 | 1,224 | 1,158 | 1,092 |
| HRB-2P | 10w buffer | 22 | Planned +486 | 200 | 178 | 372 | 350 | 598 | 576 | 554 | 532 |
| HRB-3P | 20w buffer | 7 | Needs review | 66 | 59 | 143 | 136 | 129 | 122 | 115 | 108 |
End-of-week stock
- Expected
- 1,290
- Buffer level
- 528
- Added this week
- +1,080 from an inbound lot
- Expected to land
- 21 September
Open the cell to see the lot it came from.
- Buffer weeks, not guessworkYou set how many weeks of cover you want to hold. The projection works back from it through production time, transit time and the expected arrival date, so the week a plan has to be placed is a date, not a feeling.
- Minimums where the order is placedA proposal always shows the whole shortfall, so a slow SKU is never hidden for being below a minimum. Your manufacturer’s minimums, per production SKU and for the order as a whole, are checked when the approved plan becomes a purchase order.
- A stockout does not shrink the rateDays you could not sell because stock ran low are left out of the selling rate, which is held for two weeks. If the shortage lasts, it eases down 10% a week and stops at half of what it was, so one empty shelf does not plan you into the next.
- However you pack itMake 1,000 units and pack 400 as singles and 300 as twin packs, and every Sales SKU carries its own stock, selling rate, buffer and projection, while the production run behind them stays one number.
What happens after you approve a plan.
The plan is the easy half. The part that usually lives in email and a spreadsheet is what happens between the factory and the Amazon shipment, and that is the part this holds.
- After the plan is approvedIt becomes a purchase orderThe shortfall the projection found, recorded as the order for your manufacturer. Its minimums, per production SKU and for the whole order, are checked here, before it is created.
- When the factory confirms the runYou allocate itAssign what was actually made to the shipment plans that need it, channel by channel. Anything left over stays in factory stock instead of becoming a shipment nobody planned.
- Instead of duplicating itIt maps onto what existsShipments you have already created are matched, so a lot lands on the shipment Amazon is expecting rather than creating a second one.
Every lot is a column, not a note in a file.
The inbound pipeline puts one column against each purchase order, shipment or AWD lot, and each cell is the number of boxes of that Sales SKU in that lot. A lot carries its pickup date and its estimated arrival, and says whether it is counted in your plan.
| Lot | 1 pack | 2 pack | 3 pack |
|---|---|---|---|
| Lot ARecently arrivedCompleted 23 Aug | 540 | · | · |
| Lot BIn transit, countedArrives 24 Sep | 1,080 | 270 | · |
| Lot CIn transit, countedArrives 13 Oct | 1,080 | 162 | 72 |
| New planApproved (awaiting dispatch)Arrives 24 Nov | 1,080 | 1,080 | · |
And the parts that go into it.
If you buy the components rather than a finished case, the purchase order proposes those too. Component stock is projected week by week per supplier, and where the projection turns negative the row carries the quantity to order.
- Printed carton, smallSupplier A−2,106Suggested 3,000
- Printed carton, largeSupplier A7,701No proposal
- Foil pouch, 40 countSupplier B−864Suggested 2,000
- Outer shipperSupplier B3,570No proposal
Stock does not argue an advertising decision.
The two sides are kept apart on purpose. An advertising decision is argued from the advertising record, and stock is planned here, on its own evidence.
A brief is never written because of stock. AI Analyst reads your advertising record and the relationships between your products, and it does not read Inventory at all, so no proposal arrives arguing from how much you happen to hold. Knowing what you are about to run out of is a real question — it is simply a different one, and it is answered here.
How a rule decides and what stops itWhat a Decision Brief reads
The arithmetic, without the App.
Each link in the chain is a sum you can run in a spreadsheet. Each article works one through, with a mistake that costs money and the arithmetic that finds it.
Plan the run before you need it.
Set the cover you want to hold, and the week you have to place an order stops being a feeling and becomes a date.