China Golden Week Closes in 22 Days: Turning 9.9 Sell-Through Into Your 11.11 Restock Decision

If you sell physical goods and any part of your supply chain runs through China, the stock you will have available on 11.11 is being decided this fortnight — not in October. Chinese factories typically stop accepting new production runs ahead of the 1 October National Day holiday, and with Mid-Autumn Festival also falling in late September this year, the practical cut-off for placing a purchase order that still lands in Singapore before 11.11 is somewhere between 20 and 26 September. That gives you roughly two weeks from today. The single most useful input to that decision is the sell-through data your 9.9 sale is generating over the next 48 hours, which means the reorder call has to happen before your payout reconciliation is finished, not after.

Why does 9.9 sell-through decide your 11.11 stock position?

9.9 is the cleanest demand signal a Singapore SME gets before the year-end peak. It compresses several weeks of normal demand into two days, under discount conditions similar to 11.11, across the same channels and roughly the same customer base. Nothing else in your calendar gives you that.

The mistake most operators make is treating 9.9 as a revenue event to be reconciled and 11.11 as a separate planning exercise starting in October. By October the useful part of the signal has expired, because the supply chain that would act on it is closed. The sequence that actually works is: capture sell-through velocity during 9.9, convert it into an order quantity within seven to ten days, place the PO before the holiday shutdown, and reconcile the money afterwards. Reconciliation tells you what you earned. Sell-through tells you what to buy. They are different jobs with different deadlines, and only one of them is time-critical right now.

How much lead time do you actually have before Golden Week?

Work backwards from 11.11 rather than forwards from today. A typical Singapore SME importing from a Chinese supplier is looking at production time of 15 to 30 days depending on whether the SKU is stock or customised, plus 5 to 10 days by sea freight from a South China port to Singapore, plus 3 to 5 days for customs clearance and inbound receiving, plus 2 to 3 days to get goods listed, photographed where needed, and pushed live to your marketplace listings.

That is 25 to 48 days of runway before 11.11, and it assumes nothing goes wrong. Then subtract the holiday period itself, during which nothing moves and email responses slow to a crawl. The arithmetic is unforgiving: a PO placed on 5 October for an item with a 25-day production run does not arrive in time to sell on 11.11. A PO placed on 22 September, with the same supplier and the same item, does.

Air freight buys back roughly two weeks but usually destroys the margin on anything low-value and bulky. Treat it as an exception path for your top three revenue SKUs, not as a plan.

Which 9.9 numbers should trigger a reorder?

Three signals are worth acting on, and everything else can wait for the full post-mortem.

Days of cover at peak velocity. Take units sold per day during 9.9, not your monthly average, and divide current on-hand stock by that number. Any SKU showing fewer than 20 days of cover at peak velocity is a reorder candidate, because 11.11 will run at similar or higher velocity and your replenishment window closes in two weeks.

Stock-outs during the sale window. Every hour a listing sat at zero stock is demand you never measured. Pull the timestamps and treat a SKU that sold out in the first six hours as under-forecast, not as a success. The true 11.11 quantity for that item is higher than its 9.9 units sold, sometimes materially so.

Return-adjusted sell-through. Marketplace returns on discount events commonly run higher than baseline. If your fashion or accessories categories historically return at 15 to 20 percent, reordering against gross 9.9 units will overbuy. Apply your own historical return rate by category before setting the quantity — and if you do not know that rate by category, that is itself a finding for your Q4 fix list.

What does automated reorder logic look like for a small team?

The goal is not a forecasting engine. It is removing the dependency on somebody remembering to open a spreadsheet during the busiest fortnight of the quarter.

A workable build for most SMEs is a scheduled job that reads sales and stock data from your marketplace and ERP each morning, computes days-of-cover per SKU using a trailing seven-day velocity with a peak multiplier, flags anything below threshold, and pushes the list to a WhatsApp group or email at a fixed time. Supplier lead time lives as a field per SKU, not as knowledge in one person's head, so the threshold adjusts automatically for the items that take 40 days versus the ones that take 10.

Two refinements matter more than sophistication. First, the alert should include the suggested order quantity and the supplier's name and last quoted price, so the person receiving it can act without opening three systems. Second, it should list what it deliberately excluded — discontinued SKUs, items with open POs already in transit — because a reorder alert that silently drops items trains people to distrust it.

This is a two-to-three day build against most marketplace APIs, and it keeps working for 10.10, 11.11, 12.12, and Chinese New Year restocking after that.

What if your suppliers are not in China?

The holiday closure is specific, but the underlying discipline is not. Malaysian, Indonesian, Vietnamese and local suppliers all have capacity ceilings that bind hardest in Q4, when every one of their customers is also restocking for the same peak. If you are a distributor or F&B operator sourcing regionally, the equivalent question is: what is my supplier's committed capacity for October, and have I claimed a slot in it?

Ask now, in writing, and get a confirmed production or delivery slot rather than a verbal assurance. Suppliers allocate to whoever asks first, and the ones who ask first are the ones who ran the numbers the week of 9.9.

What should you do in the next 72 hours?

Keep it to four things. Export SKU-level units sold and hourly stock levels from every channel before the data ages out of your dashboards. Identify every SKU that hit zero during the sale and note the hour it happened. Calculate days of cover at peak velocity for your top 20 revenue SKUs and mark anything under 20 days. Then send your supplier a single message this week asking for their last production slot before the October holiday, and hold that slot while you finalise quantities.

The reconciliation, the returns wave, and the full post-mortem all still need doing. None of them expire on 26 September. The restock decision does.

Frequently asked questions

How long do Chinese factories typically stay closed around National Day?

The public holiday runs for a week from 1 October, but the operational impact is usually longer. Many factories wind down production in the final days of September and take several days to return to full output afterwards, so plan for a two-week effective disruption rather than seven days. Confirm the specific dates with each supplier directly — smaller workshops and larger plants behave quite differently.

Should I reorder based on 9.9 numbers if my 9.9 performance was poor?

Yes, but interpret it correctly. Weak 9.9 numbers caused by listing problems, pricing errors or ad budget exhaustion are not demand signals, and reordering against them will leave you short on 11.11. Separate SKUs that underperformed for operational reasons from SKUs that genuinely lack demand before setting quantities. If you cannot tell the difference, that is the first thing to instrument before 10.10.

Is it worth building reorder automation for only 30 to 50 SKUs?

Usually yes, because the cost is a few days of build time and the failure mode it prevents is expensive. A single top-selling SKU out of stock through 11.11 typically costs more in lost revenue than the automation costs to build, and the same logic runs unchanged through 12.12 and Chinese New Year. Below roughly 20 SKUs, a well-maintained sheet with a weekly calendar reminder is genuinely adequate.