Import too much and your capital sits dead in the warehouse. Import too little and you stock out mid-campaign, losing revenue and tanking your platform ranking. Both outcomes are avoidable if you know how to size an order from real sell-through data and honest lead time.
Why gut-feel estimates keep missing
Many shop owners reorder by habit: last time I bought 200 units and sold out in a month, so this time I will buy 300 to be safe. The logic sounds reasonable but hides a few expensive traps.
Sell rate shifts with your ad spend. Last month you ran ads hard. This month you cut the budget. The sell rate is no longer the same. Or you are heading into a peak season and demand is about to double. Ordering by the old number without adjusting means ordering the wrong number.
Lead time is longer than it feels. From order confirmation to stock in your Vietnam warehouse, add up every leg: factory prep (3 to 7 days for in-stock items), China consolidation (2 to 5 days), sea freight (roughly 18 to 30 days) or air (3 to 7 days, much higher cost), plus customs and last-mile delivery (2 to 5 days). A sea-shipped order typically takes 25 to 45 days. If you think of lead time as "a few weeks," you are probably ordering late.
No buffer for variance. Sell a bit faster than expected, or have a shipment delayed a few days, and you are out of stock immediately. Without safety stock there is no room to absorb normal variation.
Sell-through rate is the foundation
Before any calculation, you need to know how many units a SKU sells per day or per week under your current operating conditions: your ad budget, your content output, the time of year.
The simplest way to calculate it: take total orders for the past 4 weeks and divide by 28. That gives you average daily sell rate. For a new SKU with no history, use the sell rate from weeks 2 and 3 after it went live (week one is usually abnormal from the initial traffic spike).
A worked example: a SKU sold 280 orders over the past 4 weeks. Sell rate = 280 divided by 28 = 10 units per day.
One important note: this rate must reflect the conditions you plan to maintain in the next import cycle. If you are about to increase ad spend or run a platform campaign, adjust up. If you are entering a slow season, adjust down.
The sizing formula
Once you have a sell rate and a lead time, the order quantity follows this logic:
Order quantity = (Daily sell rate x Lead time) + Safety stock - Current inventory
Each part explained:
- Daily sell rate x Lead time: this is how many units you will sell during the entire wait for the new shipment. You need enough on hand to sell through that whole window without a gap.
- Safety stock: the buffer for variance. A common approach is to multiply your daily sell rate by the number of days you want as a cushion, typically 7 to 14 days. For fast-moving SKUs, peak seasons, or items where going out of stock costs you platform ranking, lean toward the higher end.
- Current inventory: subtract what you already have to avoid over-ordering.
With concrete numbers: daily sell rate 10 units, lead time 35 days, safety stock 14 days (= 140 units), current stock 80 units.
Order quantity = (10 x 35) + 140 - 80 = 350 + 140 - 80 = 410 units.
This is not a perfect number. But it is grounded in real data, and you can explain exactly why you chose it.
Adjusting for seasonality
The formula above gives you the right number under normal conditions. When seasons shift, the sell rate has to shift with them.
Before a peak season (Tet, 11/11, 12/12): peak sell rate can run 2 to 4 times the baseline depending on the category. If you have data from last year, use the peak-period sell rate to size the order. If you have no history, multiply your current rate by an estimated factor (ask other sellers in the category or check if the platform provides trend data). The key point is placing the order early enough for stock to land before the season starts, not during it, because by then it is already too late to benefit.
After peak season: sell rate drops and you may carry leftover stock. Your next reorder must use the post-peak sell rate, not the peak-period number.
Short-trend items: apply the formula with more caution. A sea-freight lead time of 30 days means you place the order while the trend is climbing, but the stock arrives when the trend may be flat or declining. Consider using your current sell rate multiplied by a reduction factor, say 0.7 to 0.8, to account for the possibility that demand cools before the goods land.
The reorder point
Beyond knowing how much to order, you need to know when to place the order. Order too early and you lock up extra capital. Order too late and you stock out.
Reorder point = Daily sell rate x Lead time + Safety stock
This is the inventory level at which you must place a new order immediately. Not when stock hits zero, but early enough that the new shipment arrives before your safety stock runs out.
Using the same example: reorder point = (10 x 35) + 140 = 490 units. When stock drops to 490, place the order. Do not wait.
In practice, set an alert in your spreadsheet or inventory tool when stock hits this level. With many SKUs, doing this for every item is heavy work. At minimum, cover your top 10 best-selling SKUs. They drive most of your revenue, and stocking out on them hurts most.
Handling forecast error
Forecasts always miss. The goal is not perfect accuracy but keeping the error within a range you can absorb.
Small miss (selling 10 to 20 percent faster than forecast): your safety stock handles it. No urgent action needed, but note it and use the faster rate when you recalculate next time.
Large miss (selling 30 percent or more faster than forecast): revisit the sell rate and find out why. If the cause is temporary (a one-off viral video, a one-day flash sale), it does not change your base rate. If it is durable (you found a reliable creator, the platform is consistently showing you more traffic), update your base sell rate and order more.
Selling slower than forecast: this usually happens when ad spend drops or the platform algorithm shifts. Do not reorder while sales are slow. Measure your real sell rate over 2 to 4 weeks before deciding on the next import.
Build accuracy over time
The formula is only as good as the data you put in. After each shipment, record: units ordered, the date you ran out (or inventory hit its lowest point), and actual sell rate versus forecast. After 5 to 10 import cycles you will know whether your sell rate tends to run high or low, and whether a specific supplier or route consistently runs longer than quoted.
A single column in a purchase order spreadsheet is enough to start. Over several cycles your forecasts tighten because you are calibrating against real outcomes, not estimates.
Bottom line
Sizing imports correctly requires three inputs: sell rate from the past 4 weeks, honest lead time with every leg counted, and safety stock covering 7 to 14 days of variance. Combine them in the formula, subtract current inventory, and you have a number you can defend. Start with your 5 to 10 most important SKUs, build the habit, then expand from there.