Let me start with a complaint: most sellers restock on gut feeling. Last month sold well, so send twice as much this month. Got burned by one stockout, so overstock forever after. I used to do exactly that, and over one year I paid for both mistakes before I finally sat down and did the math on FBA replenishment by weeks of supply.

FBA replenishment by weeks of supply comes down to one idea: always hold enough stock to cover the next few weeks of sales, plus a safety buffer. Weeks of supply is the ruler. Before asking how much to send, ask how much this product sells per week and how many weeks the replenishment pipeline takes.

FBA Replenishment Starts with Weekly Sales, Not Monthly Sales Divided by Four

Step one is real weekly sales. Take the last eight weeks, weight the recent four more heavily and the earlier four less, and throw out abnormal weeks first: promotion spikes and stockout weeks. If you leave them in, the number lies. I sell a home-storage item that normally moves about two hundred units a week but spiked to eight hundred during a promotion week. Counting that spike straight into the average would inflate weekly sales and leave me buried in stock.

Step two is the replenishment cycle: ordering, production, ocean freight, customs clearance, and check-in at the fulfillment center. However many weeks that whole chain takes, that is how far ahead you must act. Everything you sell during that window has to be covered by on-hand and in-transit stock. Underestimate the cycle and your listing goes dark while goods are still on the water. Overestimate it and your cash sits in boxes.

My routine is a small table with four columns: weekly sales, replenishment cycle in weeks, on-hand units, and in-transit units. I update it weekly. It takes ten minutes, and it blocks ninety percent of gut-feel decisions.

The FBA Replenishment by Weeks of Supply Formula: Target Weeks Minus What You Already Have

The formula itself is simple, and here is my version. First set a target weeks of supply: how many weeks of stock you want on hand at all times. New or volatile products get a higher target; stable old products get a lower one. Then calculate:

| Item | How to calculate | | Target stock | Weekly sales × target weeks | | Stock already held | On-hand units + in-transit units | | This shipment | Target stock − stock already held |

An example. A product sells two hundred units a week, and I want six weeks of cover: that is twelve hundred units. With three hundred on hand and four hundred in transit, I already hold seven hundred, so this shipment should be five hundred. If the result is negative, do not ship. If it is zero, do not force a shipment either.

Here is a peak-season version of the same math. The same product sells two hundred units a week normally, but before peak I estimate three hundred fifty a week with a six-week target, which gives a target stock of twenty-one hundred units. With one thousand units on hand and in transit combined, this shipment should be eleven hundred. When peak ends, I switch the weekly sales and target weeks straight back to normal. The method only works if the numbers follow the season instead of staying frozen all year.

Here is the counterintuitive part: more stock does not mean safer. Many sellers think piling up inventory is insurance, but overstock costs real money. Storage fees accrue monthly, slow-moving stock gets more expensive the longer it sits, and cash tied up in boxes cannot fund your next winning product. A stockout costs you ranking for a while. A cash crunch can cost you the business.

There is also a trade-off to weigh. Bigger shipments spread freight cost thinner but raise inventory risk. Smaller, frequent shipments stay flexible but cost more per unit in head haul. My rule: bulky or heavy items ship smaller and more often; small, light items can go in bigger batches. Different size and weight, different appetite for risk.

Pitfalls That Break FBA Replenishment by Weeks of Supply

First, promotions and seasonality need their own math. A product selling two hundred units a week in normal times can double in peak season, and restocking on the normal target guarantees a stockout. Before peak, I build a separate table and raise the target by a notch, then bring it back down when the season ends.

Second, platform inventory policies change. Inbound limits and storage fee rules get adjusted from time to time, so always confirm the current numbers and thresholds against Amazon's latest policy rather than last year's screenshot in a seller group. I learned this the hard way: I staged a shipment under old rules, the rules had changed, and the goods sat waiting while I paid extra storage.

One more habit that saves me real money: every week, on the same day, I copy the in-transit number from the platform dashboard next to my own table's number and reconcile them. Once I caught a shipment that had already been checked in and made available while the dashboard still showed it in transit. Without that check I would have shipped four hundred extra units I did not need. And for a brand-new product with no sales history, I borrow a comparable older product's weekly sales, take seventy percent of it as the estimate, and replace it with real numbers after four weeks. Better to understock a new listing slightly than to bury cash in it.

Third, in-transit stock is the easiest number to get wrong. Goods already shipped, goods arrived at port but not yet checked in, and goods checked in but not yet available are three different states. Mix them together and the formula spits out a wrong shipment quantity. Reconcile weekly, and update the table the moment a status changes. Never work from memory.

Honestly, this method is not magic. It just replaces feeling with numbers. Weekly sales, replenishment cycle, target weeks: get those three numbers right and the restock quantity falls out on its own. Since I switched from guessing to this table, my stockouts dropped by more than half and freed-up cash went into new products. This trick genuinely works for me, tested in my own store.

By Ah K, E-commerce Seller (5 yrs)