Reorder Point on a Long Lane: The Four Inputs That Set It

Most sellers set a reorder point from one number: how long the goods take to arrive. On a China-to-door lane that is the least variable part of the problem. Arithmetic that survives a long lane needs a second number — how much that arrival time moves — and a decision about how often you accept being short.

Key facts

• The form is three terms. Reorder point = expected demand during lead time + a variability buffer, usually written s = x̄L + kσ_L, where k comes from the service level you choose.
• Lead time is a range. Air bands and sea bands here are estimates subject to carrier and seasonal variation, not fixed numbers.
• Buffer is a commercial choice. Cycle service level is how often you accept being short while a replenishment is in the air; it is not a setting someone else can pick for you.

What a reorder point is, stated as arithmetic

Write it on a card: multiply average daily demand by lead time in days, then add a buffer proportional to how much demand or lead time fluctuates. That factor is not a rounding habit — it comes from a chosen service level, the share of replenishment cycles you want to finish without running out.

One consequence surfaces late: a reorder point is not a forecast — it says when to order, not how much.

Why a long lane breaks the spreadsheet assumption

Short, frequent lanes let you treat lead time as constant and put uncertainty in demand. Long lanes reverse that. A booking can sit a day for a missing document, miss a departure, or wait behind a peak period; a supplier delivery can slip against a holiday calendar your plan does not contain — in 2026, Mid-Autumn on 25-27 September and the national holiday period of 1-7 October, when supplier output, receiving capacity and outbound space move together.

On sea the constraint differs in kind. The published estimates — 18-25 working days to the US West Coast, 28-38 working days to Europe — are approximations, not promises; what moves a reorder point is departures per week. A weekly sailing adds up to seven days to the tail of any delay, which no demand smoothing absorbs.

The four inputs, and where each one comes from

Table: what has to be true before the arithmetic means anything.

Most of those rows belong to you, including the service factor. That is why a forwarder cannot set your reorder point honestly: we can say what a lane tends to take, but the demand half and the risk half are properties of your business.

Worked example — illustrative numbers, not a promise

Figures use an invented SKU.

Take a product selling 30 units a day. On the standard air channel the estimate is typically 5-9 working days, so a midpoint of seven days gives expected lead-time demand of 210 units. That first term is easy. The second is the buffer: if actual door-to-door time ranged from six to eleven working days across a season — a missed departure, a document hold, a holiday week — the exposure at the top of that range is four more days of demand, roughly 120 units, before a replacement can be ordered.

Now compare two service choices. A seller who accepts being short on one cycle in ten sets the buffer near half that exposure and keeps turnover high. A seller protecting a peak window sets it at the full exposure and carries dead units all year. Same lane, same arithmetic, a different answer — and the difference is not a warehouse decision.

What changes the shape is the lane, not the buffer. Move that SKU to an estimated 28-38 working-day sea transit to Europe and the lead-time term triples before variability is added. Holding more stock to compensate is paying rent on a longer pipeline; the alternative is a faster band for the months that matter, chosen while the lane still has space.

Cycle service level is a decision, not a setting

Raising k buys availability with stock that sits still — storage time, and eventually a slow-moving line occupying a slot someone else wanted. Lowering it turns one missed departure into an out-of-stock page. The useful question is which failure costs more for this line: an empty listing in a peak week, or a pallet of unsold units after it.

Two operational facts bound it. A replenishment instruction reaching us by 16:00 China time is picked, packed and released to the line the same day; a later one joins the next working day. A stock discrepancy runs on an hourly clock — response inside two hours, resolution inside twenty-four — so the count behind your reorder point is auditable rather than assumed.

Five questions before you set a number

• Is my lead time a band or its shortest day? Built on one good week, the point is fiction.
• What is my forecast error, not just my average? The buffer is sized against the miss.
• How many departures does this lane offer per week? It outweighs most of the formula.
• Which holiday calendar sits inside my lead time? Origin weeks and destination intake deadlines differ.
• Do I pay for availability in storage or in stockouts? Answer before choosing k.

Send the per-SKU history, the band you can book and the peak window you are protecting, and we will tell you which term in the formula is doing the damage: tony@sbt56.com, or the enquiry form at Contact. Related: The China holiday calendar, Dock to available, all frequently asked questions.

Does a longer transit band mean proportionally more safety stock?

Not proportionally, and that is the trap. Expected lead-time demand grows roughly with the band; the buffer grows with variability over that window, which moves more slowly. A lane twice as long is not twice the buffer — it is more stock in transit plus a buffer sized against a wider range of arrival days, which people under-buffer.

Where should the buffer physically sit?

Inside China, at the site that receives that class of goods: parcel stock at Dongguan, palletised and crated cargo at Shenzhen, Yangtze-delta consolidation at Suzhou — and a lane departing Guangdong must arrive in Guangdong. Overseas, our points are delivery and clearance touchpoints, not return or inspection stations, so a buffer held there is a different decision with its own limits.

Can you run the reorder calculation for me?

We supply the two numbers we own: the counted quantity with the date it was counted, and the estimated transit band for the lane you are booking. The demand average, the forecast error and the service level are yours, so the multiplication is yours too. What we will tell you is when a lane's realistic range has moved enough that a point set months ago no longer holds.

Operational guidance published 31 October 2026 (2026-10-31) by FulfillDock by SBT, operating 24,000 m² of warehousing across Shenzhen, Dongguan and Suzhou under a company registration reading Shenzhen Shibotong International Logistics Co., Ltd. The form s = x̄L + kσ_L is standard inventory-control material from a published university logistics lecture; the worked figures are illustrative for an invented SKU, not a quotation or a delivery promise. All transit bands here are estimates, subject to carrier and season.