Red Sea Rerouting — The Planning Math Nobody Shows You

First, the GCC-specific trap

I want to start here because it is where I see planners in this region get it wrong most often.

Asia to Jebel Ali does not transit the Suez Canal. A container moving from Shanghai or Ningbo to Dubai crosses the Indian Ocean and enters the Gulf directly. So when Red Sea transits became unsafe, a lot of GCC planners looked at their primary inbound lane, confirmed it did not touch Bab el-Mandeb, and concluded they were insulated.

They were not, for four reasons.

Europe and Mediterranean inbound does transit it. Any lane from Northern Europe, Italy, Turkey or North Africa into the Gulf runs through Suez and Bab el-Mandeb. For distributors carrying European brands — food, personal care, automotive parts, machinery — that is a materially affected lane even when the Asia lane is clean.

Transhipment volumes move. Jebel Ali is a hub, not just a destination. Cargo that would have been relayed through the region to East Africa, the Red Sea ports or the Eastern Mediterranean gets re-planned onto different services. Hub throughput, yard density and berth availability all shift, and dwell times move with them.

Vessel capacity is a global pool. Sailing the long way around Africa absorbs ships. The same volume of trade requires more vessels to move it, which tightens capacity on every lane simultaneously — including yours. This is why rates rose on routes that never went near the Red Sea.

Your suppliers' inputs may transit it even when your finished goods do not. A supplier in Asia importing European components has your lead time embedded in their exposure.

The arithmetic that matters

Adding fourteen days to a lane does three separate things. They are usually discussed as one and they should not be.

1. Pipeline inventory — a permanent working capital transfer

Pipeline (or in-transit) inventory is simply demand rate multiplied by lead time. Extend the lead time and you permanently increase the amount of your money floating on the ocean.

This is not a one-off cost. It is a level shift in how much cash the same business ties up to run at the same service level. It never comes back until the lane shortens again. In my experience this is the number that gets a CFO's attention, because it is the one that appears on the balance sheet rather than in a freight variance line.

2. Safety stock — and why the variance term dominates

The safety stock formula that accounts for lead time variability is where the real damage shows up:

Look at the two terms under the root. The first scales with lead time linearly. The second scales with the square of daily demand multiplied by lead time variance. For any high-volume item, the second term dominates — which means your safety stock is driven far more by how unpredictable your lead time is than by how long it is.

Rerouting worsens both, but it worsens variance disproportionately. A settled lane through Suez had a tight, well-understood distribution. A rerouted lane has schedule reliability problems, port omissions, blank sailings and rolled containers stacked on top of the longer base transit. The mean moves by 54%; the standard deviation often moves by considerably more.

Run those two columns through the formula and the safety stock requirement roughly doubles — with no change whatsoever in demand behaviour. That is the point worth internalising: your inventory went up because the ocean got less predictable, not because your customers changed.

3. Forecast horizon — you are now committing blinder

The least discussed consequence. If you order 40 days out instead of 26, you are making a quantity commitment against a forecast that is two weeks further into the future, and forecast error grows with horizon. So the rerouting has quietly degraded your forecast accuracy at the point of decision, even though your model has not changed at all. Any MAPE you report against a shorter horizon is now flattering.

What to actually do

Re-baseline, do not patch. The most common response I see is adding a flat fortnight to the planning parameter. That fixes the mean and ignores the variance, which is the term doing the damage. Re-measure the actual distribution of your delivered lead times over the last two quarters and re-derive both parameters from data.

Recalculate service level as a decision, not a default. If holding 95% now costs double the buffer it used to, that is a commercial choice, not a planning setting. Some SKUs deserve 98%. Many C-class items were never worth 95% in the first place. This disruption is a reason to segment properly, which most portfolios have been putting off anyway.

Separate the structural change from the noise. Some of the added time is structural — the route is genuinely longer. Some is congestion that will ease. Model them separately, because the first should change your standing parameters and the second should not.

Push the decision point earlier, not the buffer higher. Longer lead times can sometimes be absorbed by earlier commitment on the predictable portion of demand — base volume — while keeping flexibility for the variable portion. That is usually cheaper than buffering the whole line.

Where this breaks

The formula assumes normality and independence. Real lead time distributions are right-skewed — things go late far more dramatically than they go early — and disruptions are correlated across shipments rather than independent. The standard formula will therefore understate the buffer needed during an active disruption. Treat its output as a floor.

Averages hide the thing that hurts you. A lane averaging 40 days with occasional 65-day outliers is a completely different planning problem from a reliable 40-day lane, and both look identical in a mean. Track the 90th percentile alongside the average; that is the number your stockouts actually come from.

The situation is not static. Any figure in this article is a worked illustration, not a current market quote. The method survives; the numbers need to be yours, measured this quarter.