Supply Chain and Logistics: The Ocean Is Cheap, the Doorstep Is Not

Ship a container of goods from a factory in East Asia to a port in Europe and the cost per item can be startlingly small — often a few rupees for something that will sell for hundreds.

Then move those same goods the final eight kilometres, from a local depot to somebody's front door, and that leg can cost more than the ocean crossing did.

This is the single most counterintuitive fact about physical logistics, and understanding why explains most of what happens in the field: distance is cheap, and handling is expensive.

A ship carrying twenty thousand containers is loaded once, sails once, and unloads once. A van delivering to eighty homes stops eighty times, and each stop needs a person to find the address, park, walk, and hand something over. Ocean freight is a machine problem, solved by scale. The last mile is a labour problem, and labour does not get cheaper by adding more of it.

The shape of the cost

Once you see logistics as a series of handling events rather than a distance covered, several things fall into place.

Consolidation is the core technique. Every improvement in freight over the past century has essentially been a way of handling things fewer times: standardised containers so goods are not repacked between ship, train and lorry; pallets so a forklift moves fifty boxes at once; hubs where parcels are sorted in bulk rather than routed individually.

Density beats distance. Delivering a hundred parcels in one neighbourhood is far cheaper per parcel than delivering a hundred spread across a region, even if total kilometres are similar. This is why delivery services expand area by area rather than everywhere at once, and why rural service is genuinely expensive rather than merely neglected.

Volume often matters more than weight. A van fills up with bulk long before it reaches its weight limit for most consumer goods. This is why compressing a product — vacuum-packing something bulky, or designing furniture to ship flat — can transform the economics of a business without changing the product's function at all. It is not a shipping detail; it is sometimes the whole strategy.

The general principle is a direct application of the theory of constraints: find the stage that limits throughput and attack that, rather than optimising the parts that are already fast. In most physical distribution, the constraint is the number of times a human being has to touch the item.

Efficiency and resilience come from the same budget

For several decades the dominant idea in supply chain management was to hold as little inventory as possible. The reasoning is sound: stock sitting in a warehouse is money that has been spent and not yet earned back, plus storage, insurance, and the risk of it becoming obsolete.

Running lean genuinely works, and it freed up enormous amounts of capital across many industries. But it has a cost that only appears under stress, and it is worth stating plainly rather than as hindsight.

Inventory is a shock absorber. When a supplier is late, a buffer means production continues. Remove the buffer and any delay upstream stops everything downstream immediately. The saving is collected steadily over years; the cost arrives all at once.

This is the trade that a great many organisations discovered during recent global disruptions, and the honest framing is not that lean was a mistake. It was a rational response to a long period of stable, predictable supply. The error was treating that stability as a permanent feature of the world rather than a condition that happened to hold — which is exactly the pattern described in black swan events, where a long calm justifies removing the very buffers that would absorb a shock.

The related failure is the bullwhip effect, where small changes in end demand amplify into large swings upstream. Lean systems make this worse, because there is no inventory anywhere to dampen the signal.

Where buffers actually belong

The useful conclusion is not "hold more stock." Holding stock everywhere is how you end up with capital tied up in the wrong things while still being short of what you need.

Buffers belong where uncertainty and lead time are both high, and specifically at or just before the constraint. A cheap component with a six-month lead time from a single supplier deserves a deep buffer. An expensive item available locally within two days does not.

Three other things reduce fragility more cheaply than blanket inventory:

Shorten the information delay. Most supply chain volatility is caused by decisions made on stale information. Sharing actual point-of-sale data upstream, rather than passing orders down a chain of guesses, does more to stabilise a network than extra stock does.

Have a second source for critical inputs, even at slightly higher cost. The premium is insurance, and — like all insurance — it looks wasteful until it isn't.

Know your dependencies more than one layer deep. Most companies know their suppliers. Far fewer know their suppliers' suppliers, which is where a surprising share of single points of failure sit. A component you buy from three vendors is not diversified if all three buy the same part from one factory.

The last mile remains the hardest part, and it is where most current effort is going — pickup points that convert eighty stops into one, route optimisation, delivery windows that reduce failed attempts. None of it changes the underlying fact.

Getting something most of the way across the world is a solved problem. Getting it up the stairs is not.