Resilient supply chains: beyond just-in-time

When floods in Rio Grande do Sul interrupted roads, paralyzed ports and destroyed warehouses in May 2024, many executives had the same thought: an extreme, isolated event, unlikely to be repeated anytime soon. This reading is comforting and mistaken. What Rio Grande do Sul demonstrated was not the fragility of a specific region — it was the structural fragility of chains built on the premise that the physical environment is stable. This premise is no longer true.

How just-in-time became a trap

Just-in-time logic was developed by Toyota in the 1950s in a specific context: limited physical space, scarce capital and a controlled production environment. The core idea—eliminating idle inventory by reducing the time between order and delivery—produced real efficiency gains and became the dominant model for decades to come.

The problem is that JIT has been optimized for cost. Under normal conditions, minimizing inventory and maximizing speed is great. When the external environment becomes unpredictable — by a virus, by a cyclone, by a drought that reduces the navigability of a river used as a transportation route — efficiency becomes vulnerability. Any interruption propagates to the final product without damping.

The pandemic was JIT's first major stress test on a global scale. Chains of semiconductors, pharmaceuticals, and medical equipment crashed in ways that took years to recover. The lesson many companies took away was narrow: diversify chip suppliers, bring production of critical items closer together. The broader lesson—that optimizing for cost needs to incorporate optimization for resilience—took a backseat as costs began to pressure margins again.

The climate as a shock that doesn't go away

The difference between the pandemic shock and the climate shock is that the pandemic was a discrete event. The chains suffered, adapted and resumed operations. Extreme weather events don't work like that: they become more frequent, more intense and more geographically dispersed. What was considered a 100-year event began to occur with much greater frequency in several regions.

Risk operates in two simultaneous dimensions. The first is direct physics: infrastructure damaged by floods, ports paralyzed by hurricanes, roads cut off by landslides. The second is inputs: extreme heat reduces agricultural productivity, drought compromises the navigability of rivers used as transport routes, water stress limits industrial capacity. Both affect the same result: product availability and delivery cost.

Geographic concentration amplifies the problem. Rice production in the South of Brazil, shoe manufacturing in Vale dos Sinos, soybeans in the Center-West: when climate events hit these concentrated regions, the effect propagates throughout entire chains that depend on these supplies.

Climate risk mapping in the chain

The starting point is to know where the risk is. Most companies have reasonable visibility into their direct suppliers (Tier 1) and very limited visibility into their suppliers' suppliers (Tier 2 and Tier 3) — which are often where the most vulnerable concentration points are.

Chain climate risk mapping means crossing the geographic location of suppliers, production facilities, transport routes and storage points with physical risk data — flood models, drought maps, frequency projections of extreme events. This crossing reveals which nodes in the chain are in high exposure zones and which critical dependencies pass through these zones.

Tools that combine geospatial physical risk data with chain mapping already exist and have been adopted by global retailers, automakers and consumer goods companies. The cost of access has dropped significantly, making the exercise viable for medium-sized companies.

The real cost of resilience

The most common argument against investing in chain resilience is the cost: maintaining safety stocks immobilizes capital, duplicating suppliers increases procurement costs, diversifying routes increases logistics costs. The argument is correct — resilience has a cost. What the argument ignores is the cost of disruption.

The real cost of an outage goes beyond lost sales. Includes emergency replacement (air freight, spot suppliers at premium prices), loss of customers who migrated to substitutes, factories stopped waiting for input and market positions given to less affected competitors. Added together, these costs often exceed the investment that would have been necessary to avoid the disruption.

The most efficient way to build resilience is criticality analysis: which inputs, if missing, stop the entire operation, and which have substitutes or can be postponed without immediate impact. For critical items, buffering or source redundancy is justifiable. For items of low criticality and easy replacement, the cost is probably not justified.

Strategic redesign: what changes in practice

Redesign involves decisions across three horizons that should not be mixed into a single program — this is a common mistake that creates paralysis.

In the short term, the work is visibility and contingency: completing risk mapping, identifying the three to five most critical nodes, and defining specific contingency plans — not generic "business continuity" plans, but concrete responses to specific scenarios. Which alternative supplier would be activated, with what lead time, at what incremental cost and with what qualification process.

In the medium term, the central decision is where to accept the cost of permanent redundancy versus where to optimize for speed of response. Some companies will find that maintaining a second qualified supplier in a different region costs less than an uncovered outage. Others will find that the spot market is efficient enough that permanent redundancy is not worth it.