Expedited freight vs waiting: the break-even math

One inequality decides every expedite: the dark days you'd prevent, times the profit those days would have earned, against the all-in premium. Left side bigger — pay. Right side bigger — wait. Merchants who get this wrong aren't bad at arithmetic; they feed the arithmetic the wrong inputs: a sales velocity read after the shelf went dark, a premium that's only the freight line, and a loss estimate that assumes every customer waits around for a restock.

This article builds the worksheet once, on a realistic SKU, and shows where the decision flips. The triage side — which SKUs to rescue first when several are burning at once, and how a partial air top-up compares to a full expedite per dark day removed — lives in the emergency restock playbook; this piece is the formal worksheet that comparison runs on. And the broad mode question, air versus sea for your whole replenishment model, is the air versus sea tradeoff at the shipment level. This is narrower and more urgent: one SKU, one crisis, expedite or wait.

The inputs, and the velocity trap

Three numbers drive the decision, and the first one is where most analyses break.

In-stock velocity. A SKU sitting at zero records zero sales — Shopify reports recorded orders, not suppressed demand (Shopify Help Center, accessed November 2026) — so the velocity you can see during a gap is a lie in the useful direction: it says the SKU stopped needing customers. Use the trailing 4 weeks of in-stock days only for the daily rate. If your records blur stock and sales dates, approximate from the 4 weeks before the shelf emptied.

Profit per unit. Price minus unit cost minus inbound freight per unit minus selling fees. Not contribution margin in the textbook sense — the cash the unit actually adds when it sells.

Dark days projected. Days from projected zero until sellable stock lands, measured on your real lead time — not the supplier's quote.

Worked example, bike lights, all numbers illustrative:

SKU BIKE-LT-FR-02 (front light): price $34.50, unit cost $11.20, inbound freight $1.60/unit, platform fees $1.90 → $19.80 profit per unit In-stock velocity: 9 units/day (trailing 4 weeks) On hand: 54 → hits zero in 6 days; the ocean PO lands day 34 → dark window: 28 days

The break-even worksheet

The break-even form is the durable takeaway: break-even days = all-in premium ÷ adjusted daily profit at risk. Expedite only when the projected dark window exceeds it.

Now watch the same SKU flip. Same product, same premium, same 28-day window — but a tail variant selling 4/day instead of 9:

Daily at risk: 4 × 19.80 = 79.20 → adjusted: 79.20 × 0.55 = $43.56 Loss if wait: 28 × 43.56 = $1,220 estimated — under the $1,800 estimated premium Break-even: 1,800 ÷ 43.56 = 41.3 days — the window is 28, so wait

Nothing about the SKU changed except how fast it sells, and the decision reversed. This is why expediting "the catalog" is never a policy — it's a per-SKU division.

The recovery factor: not every lost sale is lost

The raw daily-at-risk number overstates the loss, because customers don't all wait. The best-known data on what actually happens comes from a 2002 out-of-stock study commissioned by the Grocery Manufacturers of America for the Coca-Cola Retailing Research Council: of shoppers who hit an empty shelf, 31% bought the item at another store, 26% bought the brand elsewhere, 19% substituted, 15% delayed the purchase, and 9% bought nothing. Two decades old and grocery-based, so treat it as a shape, not a constant — but the shape is the point: a meaningful share of demand survives the gap.

Translate it into the recovery factor:

• Delayers (~15%) mostly come back once stock lands — not at risk.
• Substitutors (~19%) stay with you if you have a sibling variant in stock — your other light model catches them. No sibling, no catch.
• The rest (~31% + 26% + 9%) is genuinely gone — bought elsewhere or not at all.

That yields two defensible bookends: roughly 0.66 at-risk for a hero SKU with no substitute on your own shelf, and roughly 0.47 when an in-stock sibling absorbs the substitutors. The worksheet uses 0.55 as the illustrative middle for a catalog with partial substitutes. Pick your SKU's side deliberately — using the no-substitute factor for a SKU that has three variants in stock is how merchants talk themselves into air freight the math didn't approve.

The premium is more than the freight line

The freight quote is the largest line and never the whole premium. Air-freighting a top-up creates a second shipment, and second shipments carry their own fixed stack — in the worksheet: origin pickup $90, ISF and customs entry $150, destination delivery $90, all illustrative. Skip those and you've understated the premium by ~20%, which on a borderline decision is the difference between expedite and wait.

There's also a per-unit effect when you expedite the pipeline itself. The worksheet's $1,800 estimated premium spread over 900 air-freighted units adds $2.00 per unit (illustrative) to landed cost — $19.80 profit becomes $17.80. On a $19.80-margin hero that's survivable; on a $4-margin tail SKU it removes half the profit of every unit in the pipeline. Margin compression is a second reason the tail waits even when the window is long.

One pricing note for the freight line itself: air rates are quoted per kilogram on the greater of actual or volumetric weight (volumetric divisor 6,000 cm³/kg is the common standard — Freightos's glossary documents it, accessed November 2026), and they move weekly. Light, bulky cartons pay for air they don't weigh — another reason to expedite units, not cartons.

Expedite the days, not the PO

The full-pipeline expedite from the worksheet costs $1,800 estimated and prevents 28 dark days. But the decision has a cheaper cousin: air a top-up sized to bridge the gap, while the ocean PO sails as planned. The sizing formula:

top-up units = velocity × (ocean ETA − top-up ETA) + safety days × velocity

Whether the top-up pays depends entirely on how far away the ocean box is — run it both ways:

Ocean lands day 34 (the worksheet case). The top-up arrives day 13, so it prevents only days 6→13: 7 dark days × $98.01 = $686 estimated, against a ~$750 estimated premium once the fixed stack is counted (60 kg ≈ $420 freight + $330 fixed). Break-even demands about 7.7 prevented days ($750 estimated premium ÷ $98.01 per dark day); the top-up delivers 7. Fails — it mostly pays for its own fixed stack.

Ocean lands day 54. Dark window becomes 48 days, the top-up prevents 41 of them: $4,018 estimated protected against a ~$1,030 estimated premium (100 kg ≈ $700 freight + $330 fixed). Clears decisively — and for $770 estimated less than expediting the whole pipeline, with the ocean stock still coming as backup.

The pattern is the article in one paragraph: a top-up pays when the ocean box is far, and fails when it's near — because the fixed stack is constant while the prevented days depend on the gap it bridges. Always compute the prevented days the top-up actually covers (arrival to ocean ETA), not the whole projected window. The playbook's ranked triage runs this same partial-versus-full comparison per dark day across all six rescue options; the numbers above are what that ranking looks like through this worksheet's all-in-premium lens.

When waiting is the right answer

• Tail SKUs, almost always. Thin per-unit margins make the per-unit premium fatal; break-even days stretch past any realistic window (the 4/day flip above).
• Your catalog catches the substitutions. If an in-stock sibling variant absorbs the substituting share, the recovery factor drops and the loss-if-wait shrinks with it — count your own substitution before paying to prevent it.
• The loss-if-wait is under the premium. The boring verdict, and the most common one at low velocity: $1,220 estimated against $1,800 estimated is a wait, however uncomfortable the empty shelf feels.
• The fix is buffering, not speed. A SKU that keeps arriving late enough to need air freight has a lead-time or buffer problem — sizing a deliberate buffer for the next cycle is the recurring fix, and it's bought once instead of per crisis.

The exception that overrides the math: deadline value. A launch date, a wholesale commitment, or the four weeks that decide your quarter carries value the per-unit math can't see. Treat deadline overrides as explicit judgment calls — write down what the deadline is worth before paying for it — rather than letting urgency pose as arithmetic. And remember the arithmetic above counts SKU profit only; the wider bill for a dark SKU (wasted ad spend, rank decay, subscription churn) lives in the full cost of a SKU sitting at zero, and it only pushes further toward expediting.

Common questions

When is expedited freight worth it? When the profit you'd lose during the dark window exceeds the all-in premium of the expedited shipment. In the worked example: 28 dark days × $98.01 adjusted daily profit at risk = $2,744 estimated loss, against a $1,800 estimated premium — expedite. The same SKU at 4 units/day loses $1,220 estimated, under the premium — wait.

How do I calculate the break-even for expediting? Break-even dark days = all-in premium ÷ adjusted daily profit at risk, where daily at risk = in-stock velocity × profit per unit × recovery factor. A $1,800 estimated premium against $98.01/day breaks even at 18.4 days; expedite only when the projected dark window exceeds that number.

Do customers wait for a restock? Not most of them. The 2002 Grocery Manufacturers of America study for the Coca-Cola Retailing Research Council found 31% bought elsewhere, 26% bought the brand elsewhere, 19% substituted, 15% delayed, 9% bought nothing. Roughly half to two-thirds of demand is genuinely at risk, depending on whether your own catalog offers a substitute.

Should I expedite the whole PO or a partial top-up? Depends on how far the ocean box is. Sized as velocity × (ocean ETA − top-up ETA) + safety, a top-up fails when the ocean PO lands soon (7 prevented days against a ~$750 estimated premium) and clears decisively when it's far (41 prevented days for ~$1,030 estimated). Compute the days the top-up actually covers, not the full window.

Is air freight charged by weight or volume? By weight — specifically the greater of actual or volumetric weight, with volumetric weight computed at a divisor of 6,000 cm³/kg in common practice (Freightos glossary, accessed November 2026). Bulky light cartons pay for the space they occupy, which is why expediting compact units beats expediting cartons of air.

Next step

Which SKU is closest to zero right now? Pick it and fill the worksheet above for it today — trailing in-stock velocity, profit per unit, the recovery factor that matches its substitutes, and a real dark-window estimate from its ocean ETA. If the loss-if-wait beats the premium you're quoted, the expedite decision is already made before the quote expires; if several SKUs qualify at once, the rescue-order playbook sets the sequence.