Booking factory capacity for peak season: when to reserve, and what it costs
When to reserve factory capacity before peak season: a booking calendar working back from order-by dates, deposit norms, and the cost of reserving late.
InventoryX is in development and not available today. ClearStock is the live Shopify app.
Somewhere around July, your factory's sales rep asks the question that decides next spring: which production slots do you want reserved for Q1? Answer late (say, December) and the answer comes back as a queue position behind buyers who reserved in summer, just as Chinese New Year (February 6, 2027) prepares to shut most Guangdong and Zhejiang lines from late January until well into February, materials suppliers included. Capacity, at a small factory, is queue position, and queue position is the one thing money can't buy after the line is full.
This article lays out the calendar that turns "book capacity" into actual dates: how to work back from the day stock must be sellable, what a reservation costs and what skipping it costs, and the two shutdowns that bend every date. Verifying the factory itself (can they actually produce what they claim, at what quality?) is a separate discipline, covered in the factory audit guide. Reserving the calendar comes first, though.
What "capacity" means on a production line
A factory's peak-season problem is that September and October are finite. The line runs one order at a time, and the biggest buyers (the ones doing 40-foot containers monthly) book their Q4 slots months ahead, because their purchasing teams do exactly the math in this article. When your PO arrives in late August unreserved, it doesn't jump the queue; it joins it, behind slots already promised, in the season when every buyer's urgency peaks simultaneously.
Three things stack on top of each other:
- Line time: days of machine and crew time to produce your units. The part everyone thinks of.
- Materials lead time: webbing, hardware, boxes. Components arrive on their own calendars, and component suppliers shut down for the same holidays. A factory can't start your order without materials, so a 20-day component delay pushes the whole slot.
- The queue. The gap between "PO arrives" and "production starts." In July, this can be days. In late September, weeks (illustrative; ask your factory what theirs currently is, and most will just tell you).
Reserving capacity means converting the third item from a gamble into a schedule: a PO with a production start date, confirmed in writing, with a deposit behind it.
The booking calendar, worked back from your order-by date
The math is the same as any lead-time calculation, with one twist: the lead time you plan against is the booked number, not the quoted one. Worked example, illustrative throughout: YOGA-MAT-PUR-01 sells 6.5 units/day, and you want 90 days of cover entering September. Required: 6.5 × 90 = 585 units.
Total lead time on a booked slot: production 45 days + ocean transit and clearance 35 + receiving 3 = 83 days. To be sellable September 1, 2027: order by June 10, 2027. Now the capacity twist: an unreserved PO arriving in the pre-Q4 crunch typically waits an extra 15 days in queue (illustrative), stretching effective lead time to 98 days and pushing your order-by to May 26. Reserving the slot doesn't speed up production; it moves your order-by two weeks later, which is two more weeks of real sales data before you commit to a quantity.
From November 2026, the calendar holds three windows:
| Window | What closes | When to act | Notes |
|---|---|---|---|
| Chinese New Year, Feb 6, 2027 | Factories wind down from late January; returns stretch through February, materials suppliers included | Order by mid-December 2026 for anything needed before the holiday; plan for a ~4-week dead zone | The year's biggest disruption; dates vary by province and factory (China Briefing, accessed November 2026) |
| Golden Week, Oct 1–7, 2027 | A full week plus ramp-down and ramp-up days on either side | Order by mid-August 2027 | Shorter than CNY but lands right before your peak; assume little ships Oct–early Nov that isn't already booked |
| Q4 line slots | September–October production calendars fill from July onward | Reserve by early July 2027 | This is the window this whole article exists for: deposit in hand before August |
One caveat on the deposit column: a 30% deposit with the balance before shipment is a common structure for locking slots, but terms are negotiated, not universal. Treat every norm here as a starting point for your own contract, not a published rule.
Why July for stock that sells in November: the Q4 slot produces September–October output, which is what stands on shelves through Black Friday–Cyber Monday. The demand side of that window (how much to hold, when the promo weeks start) is the territory of the BFCM inventory planning hub; this calendar exists to get the goods onto your shelves in time for it.
What reserving costs, and what not reserving costs
The reservation bill, on the illustrative yoga-mat order: a deposit of 30% × (585 × $14.20) = 30% × $8,307 ≈ $2,492 estimated cash tied up for the production cycle, plus a MOQ commitment and a price locked before peak-season material moves. That's the full cost. Ocean freight rates do climb with peak-season demand (the weekly movements are published in the Freightos Baltic Index, accessed November 2026), but a booked slot with a locked price largely insulates your unit economics from that.
Not reserving costs more, and it costs in the currency that hurts most: time. Three ways the bill arrives:
- The queue extension: 15 extra days (illustrative) of effective lead time, which either pushes your stock arrival into your peak or forces you to order earlier on worse data.
- The dark days. If the queue pushes arrival past your last unit: 6.5 units/day × 15 days = 98 units × $42 retail = $4,116 estimated in missed revenue on this SKU alone.
- The panic premium: air freight is the classic rescue, and for bulky goods it's brutal. Air-freighting 585 yoga mats (volumetric-heavy cargo) can add an estimated $3,000–$4,500 illustrative over ocean. Paying it once teaches the lesson; the reservation is the cheaper lesson.
Note the asymmetry: the deposit is cash you'd spend anyway, paid earlier. The non-reservation costs are pure waste: revenue that never happens, freight premium that buys nothing but the status quo.
CNY and Golden Week: the two shutdowns that bend every date
Chinese New Year is the deep one. Factories typically wind down from late January, and re-opening is gradual — crews return through mid-to-late February, and upstream materials suppliers follow the same rhythm, so even a factory back at full staff may wait on fabric or hardware (China Briefing, accessed November 2026). Treat late January through late February as a ~4-week dead zone every year, and place December orders with that cliff in plain view: an order confirmed December 15 either finishes before the wind-down or realistically lands in March. The full countdown arithmetic for the CNY window (order-by dates working back from the shutdown) is laid out in the Chinese New Year buying deadline guide.
Golden Week (October 1–7) is shorter but nastier in placement: it sits directly on your peak runway. The practical rules: nothing ships during the week, pre-holiday capacity goes to booked slots, and post-holiday ramp-up runs a week or more behind schedule. Inspection capacity compresses too, since third-party firms book out before both holidays (QIMA publishes guidance on scheduling inspections around CNY, accessed November 2026); a failed inspection right before a shutdown costs you weeks, not days, to re-run. The ramp-down and recovery around Golden Week have their own shutdown playbook.
What a capacity reservation actually looks like
Not a handshake. Four elements, in writing. And if the relationship runs through a manufacturer's rep, collect all four from the factory itself: the rep sells your line, but the slot confirmation must come from the floor that produces it.
- A PO with a production start date, not just a ship date: "production starts August 25" is checkable; "ships in time" is a wish.
- Written slot confirmation: an email from the factory naming the date and the line, reply kept. This is the document the whole negotiation rests on if things slip.
- The deposit, wired against the confirmed slot. No deposit, no slot; the confirmation letter without the deposit is a priority-free promise.
- Materials locked: confirmation that components for your order are ordered. A reserved line with missing webbing produces nothing.
All four live in the PO and its annex, and the discipline of keeping those terms explicit is exactly what the purchase order process guide covers. Even with all four, factories still slip: a reserved slot shifts the odds, it doesn't repeal physics. The audit helps you know which factories' confirmations are worth the paper, which is why verification and reservation are separate skills with separate checklists.
Turning the calendar into weekly order-by dates
The calendar above is one SKU. A real store runs dozens, each with its own velocity, lead time, and cover, and the arithmetic per SKU is the same working-back exercise repeated. This is the one place we'll mention what we build: InventoryX's weekly Monday Buying List turns that working-back into a per-SKU list (days of cover, suggested quantity, and an order-by date) using the manual lead times you enter per supplier. V1 doesn't auto-learn lead times, and it can't know about your reserved slot until you tell it: once the factory confirms a booked production time, enter that number as the lead time yourself, and the order-by dates inherit it. If the working-back math feels abstract, the days-of-cover guide shows the metric that drives it, and the supplier lead time guide shows how to measure the inputs accurately.
Common questions
When should I book factory capacity for Q4? Work back from your sellable date: total lead time (production + transit + receiving) plus the unreserved queue risk. For a September 1 sellable date with an 83-day booked lead time, order by June 10 — and reserve the line slot by early July, when September–October calendars fill. Deposits typically lock slots; confirmations without deposits don't.
What does it cost to reserve a factory slot? Usually a deposit — 30% of the PO with balance before shipment is a common structure (negotiated, not universal) — plus a MOQ commitment and a price locked before peak material moves. On an $8,300 estimated PO that's roughly $2,500 estimated held for the production cycle. It's cash you'd spend anyway, paid earlier; the costs of not reserving are pure waste.
How long do Chinese factories close for Chinese New Year? Treat it as a roughly four-week dead zone: wind-down from late January, gradual return through mid-to-late February, with materials suppliers on the same rhythm. For 2027, Chinese New Year falls on February 6, so December orders must either finish before the wind-down or plan on March arrival. Exact dates vary by province and factory.
Is a slot confirmation email enough to hold capacity? Only with a deposit behind it. The enforceable version is a PO naming a production start date, a written confirmation naming the line and date, the deposit wired, and components ordered. A confirmation email without the deposit is a priority-free promise: useful as evidence, useless as a schedule.
What happens if I don't reserve capacity before peak? Your PO joins the queue behind buyers who did (an extra 15 days is illustrative in the pre-Q4 crunch), which either delays stock into your peak or forces earlier orders on worse data. If stock runs out mid-wait, the rescue is air freight, which for bulky goods can add thousands estimated over ocean. The reservation converts that gamble into a date on a calendar.
Next step
Pick your one most seasonal SKU today and run the working-back math on it: sellable date, booked lead time, order-by date, and the deposit the reservation needs. Put the order-by date on the calendar now, then check the same math for your other SKUs with the days-of-cover guide and measured inputs from the supplier lead time guide, and the whole calendar assembles itself.