Chapter 1 of 4

Lead-Time Geometry and the Global Pipeline

3 sections · ~6 min read · 3 practice questions · part of Global Supply Chains

Learning objectives

  • Translate geographic distance and mode choice into inventory pipeline days.
  • Compute pipeline inventory and the implied working-capital cost across regions.
  • Diagnose where a global pipeline is exposed to schedule risk.

Distance, Modes, and Pipeline Days

In a domestic chain, a day's pipeline is roughly a day; in a global chain, the same day's pipeline is rarely the same. Pipeline days are a function of distance, mode, port handling, customs clearance, and last-mile delivery. A factory-to-customer lane of 8,000 km might run as ocean (≈ 28 days port-to-port, plus 5 days handling each side and 2 days inland), totaling roughly 40 days from order release to receipt. The same distance by air is 2 days plus handling. The implication is that 'global' is not just a label but a structural lead-time penalty. Inventory planning that treats a 40-day lane like a 10-day lane will chronically under-buy safety stock. Conversely, treating a 10-day lane like a 40-day lane will tie up unnecessary capital. The discipline is to model each lane's actual pipeline days end-to-end, refresh them as carriers and ports change, and assign safety stock accordingly. A global chain's lead-time geometry is its first design constraint: it dictates how responsive the chain can be to any given demand signal.

Pipeline Inventory and Working Capital

Average pipeline inventory equals demand rate × pipeline days (assuming continuous demand). On a lane serving 5,000 units/day with a 40-day pipeline, average in-transit inventory is 5,000 × 40 = 200,000 units. At $20 unit cost and a 9% cost of capital, the implied carrying cost per year on the pipeline is roughly 200,000 × $20 × 9% = $360,000 (full-year carrying cost on the average pipeline balance). This is the hidden cost of global scale: every pipeline day costs real money. Mode choice is therefore a working-capital decision, not only a freight decision. A lane saved is not only a freight saving but an inventory carrying cost reduction. The right comparison is total landed cost per unit delivered, including transit-time-driven inventory carrying cost. The trap is to evaluate the freight-only decision and ignore the inventory effect, which usually dominates for medium-to-high-value goods. Companies that understand this often find their global footprint has more working-capital tied up than the executive team appreciates.

Schedule Risk in Long Pipelines

Long pipelines are not just slow; they are unreliable. Variance in transit time grows with distance, with the number of handoffs, and with the number of parties whose operations are outside the buyer's control. A 40-day lane might have a standard deviation of 4 days; a 10-day lane might have 1 day. The standard deviation of demand during lead time is dominated by lead-time variance, not demand variance, on long global lanes. Schedule risk also has direction: a missed sailing cannot be accelerated, only rebuilt; a missed truck can often be caught up with expedited freight. The operational implication is that long-pipeline lanes need earlier forecasting and earlier commitment, more buffer at the receiving warehouse, and clearer communication with downstream customers. Companies that treat a global lane as if it were a domestic lane often discover their risk profile only after a major stockout. The discipline is to measure transit-time variance on every lane, use it in safety-stock sizing, and explicitly allocate forecast horizon by lane.

Worked example

Problem

A North American apparel company sources a fabric roll from a Southeast Asian mill serving 4,800 units/day average demand. The lane runs 28 days ocean port-to-port, 4 days inland handling at origin, 3 days inland at destination, with a transit-time standard deviation of 5 days. Unit cost $14, cost of capital 9%. Compute average pipeline inventory, working-capital tied up, the implied annual carrying cost on that working capital, and the implied safety stock at 95% CSL (z=1.65). Daily demand standard deviation σ_d = 240 units.

Step by step

  1. Pipeline days = 4 (origin handling) + 28 (ocean) + 3 (destination) = 35 days.
  2. Average pipeline inventory = 4,800 × 35 = 168,000 units.
  3. Working capital in pipeline = 168,000 × $14 = $2,352,000.
  4. Annual cost-of-capital on pipeline working capital = $2,352,000 × 0.09 = $211,680 per year. A useful way to hold this: each pipeline day costs 4,800 × $14 × 0.09 = $6,048 per year, and 35 days × $6,048 = $211,680. Compressing the lane by one day is worth $6,048 a year, every year.
  5. Lead-time standard deviation: σ_LT = √((σ_d × √LT)² + (μ_d × σ_LT_days)²) where μ_d = 4,800, σ_LT_days = 5, LT = 35.
  6. Demand component: σ_d × √LT = 240 × √35 ≈ 240 × 5.9161 = 1,419.86 units.
  7. Lead-time component: μ_d × σ_LT_days = 4,800 × 5 = 24,000 units. Total σ_LT = √(1,419.86² + 24,000²) = √(2,015,994 + 576,000,000) = √578,015,994 ≈ 24,042 units.
  8. Safety stock = z × σ_LT = 1.65 × 24,042 ≈ 39,669 units. In dollars ≈ 39,669 × $14 = $555,366.
  9. Total working capital: pipeline + safety stock = $2,352,000 + $555,366 = $2,907,366.

Answer. Pipeline days = 35. Average pipeline inventory = 168,000 units ($2.35M). Annual carrying cost on pipeline ≈ $211,680. Safety stock at 95% CSL ≈ 39,669 units ($555k). Combined working capital ≈ $2.91M. The lead-time component dominates: nearly all variance comes from port and carrier variability, not demand variability. Limitation: assumes independence of demand and lead-time components; in tight markets they can be weakly correlated, increasing actual σ_LT.

Practice

Work each question before opening the solution.

  1. Why does the lead-time term often dominate the demand term in safety-stock sizing for global lanes?

    Show solution for question 1

    The two terms scale differently. The demand contribution to the standard deviation is σ_d × √LT, so it grows only with the square root of lead time. The lead-time contribution is μ_d × σ_LT_days, which scales with the full mean demand rate and does not benefit from any square-root damping. On a long lane with a high-volume item, μ_d is a large number multiplying the lead-time standard deviation directly, so that term dwarfs the other. In the worked example above, the lead-time term is 24,000 units against 1,420 for demand, so demand variability contributes essentially nothing to the buffer. The practical consequence is that on such a lane, forecast accuracy work will not reduce inventory; only transit-time reliability will.

  2. What is the working-capital cost of one extra pipeline day, and in what units should it be quoted?

    Show solution for question 2

    One extra pipeline day adds one day of demand to the average in-transit inventory. The working capital added is daily demand × unit cost, and the annual cost of carrying it is that figure multiplied by the cost of capital. For a lane running 5,000 units/day at $20 with a 9% cost of capital: 5,000 × $20 = $100,000 of additional working capital, costing 0.09 × $100,000 = $9,000 per year. Quote both figures and be careful not to conflate them: $100,000 is a one-time balance-sheet effect, while $9,000 is the recurring annual charge. Writing $9,000 as a per-day figure, as is easily done, overstates the cost by a factor of 365.

  3. Why is air freight sometimes cheaper than ocean in total cost even at higher freight rates?

    Show solution for question 3

    Because air shortens the pipeline, reducing pipeline-inventory carrying cost. For high-value or highly time-sensitive goods, the inventory carrying cost saved can exceed the freight premium.