How to stress-test your supply chain: a step-by-step guide for manufacturers

01What a supply chain stress test is

A supply chain stress test is a structured "what if" exercise. You take a model of how your company buys, makes and ships, apply a shock that could realistically happen, and measure the consequences on the numbers that matter to the business: gross margin, cash tied up in inventory, lead times and the share of orders you can still deliver on time.

The idea comes from banking. After the 2008 crisis, supervisors in the United States and Europe began requiring banks to show how their capital would hold up under adverse macroeconomic scenarios. The European Banking Authority and the Federal Reserve publish those scenarios, and banks must prove they can survive them. Nobody requires a manufacturer to do the same, but the logic transfers directly: the point is not to predict the future, it is to know in advance how exposed you are and what you would do.

A stress test is different from a risk register. A risk register lists things that could go wrong and gives each a score from 1 to 5. A stress test puts a number in euros on a specific scenario and follows it through the chain: which inputs get more expensive, which deliveries arrive late, which products lose margin, and by how much.

02Why small and mid-sized manufacturers need one

Large companies have risk teams and pay for enterprise platforms that monitor thousands of suppliers. Smaller manufacturers usually have neither, yet they are often more exposed:

• Fewer suppliers per input. A company with 30 critical suppliers frequently has several inputs with a single source. One failure stops a production line.
• Thinner margins and less cash. A 3-point drop in gross margin can wipe out a year's profit for a company running at 6% EBIT.
• Less negotiating power. A small buyer is the last to get allocation during a shortage and the first to absorb a price increase.
• Decisions by instinct. In our interviews with Italian manufacturers we heard the same answer many times: when tariffs or freight rates move, decisions are made "by feel" or after paying a consultant, weeks later.

The events of the last few years have made the case. The COVID-19 port congestion of 2021, the blockage of the Suez Canal by the Ever Given in March 2021, the energy price spike in Europe in 2022 and the diversion of container ships around the Cape of Good Hope from late 2023 all hit companies that had never modelled those scenarios.

03Step 1: map what you depend on

You cannot stress-test what you have not mapped. The map does not need to be complete on day one. It needs to cover the inputs that drive most of your cost and the ones you cannot easily replace.

For each critical input, record:

Then add your own plants and warehouses, and the main customer markets. That is enough to start. Each additional detail, such as tier-2 suppliers or alternative routes, improves precision, but the first version should take an afternoon, not a quarter.

04Step 2: choose severe but plausible scenarios

A good scenario is specific, has a cause, and has a magnitude taken from history or from an announced policy. "Supply disruption" is not a scenario. "The Red Sea route is unavailable for six months and container freight from China to Northern Europe doubles" is.

Useful families of scenarios for a manufacturer:

1. Trade policy. A new tariff on an input or on your exports. Historical anchor: the US Section 232 tariffs of 2018 set 25% on steel and 10% on aluminium.
2. Chokepoints and logistics. Closure or restriction of the Suez Canal, the Strait of Hormuz, the Panama Canal or a major port. See our guide to shipping chokepoints.
3. Sanctions. A supplier, its owner or its country added to a sanctions list, forcing an immediate stop. See supplier sanctions screening.
4. Commodity and energy prices. A jump in gas, electricity, steel or polymer prices. European gas prices in 2022 reached many times their pre-crisis level.
5. Currency. A 10% move of the euro against the dollar or the renminbi when you buy in one currency and sell in another.
6. Supplier failure. Bankruptcy, fire or flood at a single-source supplier.

Run at least one scenario per family that applies to you, plus one combined scenario. Real crises rarely come alone: the 2022 energy shock came with higher freight, a stronger dollar and new sanctions at the same time.

05Step 3: trace the shock through the chain

This is where most spreadsheets break. A tariff on a component does not only raise the price of that component. It also raises the prices of everything your suppliers make with it, which reaches you weeks later through other invoices. Economists measure these indirect effects with input-output tables, which describe how much each sector buys from every other sector in every country.

For each scenario, compute three layers of impact:

• Direct cost: the tariff, freight surcharge or price increase on the inputs you buy, using your actual volumes.
• Indirect cost: the share of the shock that reaches you through suppliers' own costs. Input-output data gives a reasonable estimate by sector.
• Operational impact: longer lead times, extra inventory needed to keep the same service level, and lost sales if a single-source input stops.

Margin impact (€) = Σ inputs [ volume × (Δ direct price + Δ indirect price) ]
                  + extra inventory × cost of capital
                  + lost contribution margin from missed orders
                  − share passed on to customers

06Step 4: measure ranges, not single numbers

Every input to a stress test is uncertain: how long the closure lasts, how much freight rates rise, how much of the cost your customers accept. Treating each as a single number gives a precise answer to the wrong question.

Monte Carlo simulation solves this by running the scenario thousands of times, each time drawing the uncertain inputs from a realistic range. The result is a distribution of outcomes. Two numbers summarise it well:

If the median impact is a 2-point loss of gross margin but the bad case is 9 points, the right response is very different from a case where both are close to 2.

07Step 5: rank the responses

A stress test that ends with "we would lose €600,000" is only half done. The useful output is a ranked list of what you can do about it, each response simulated against the same scenario:

Rank them by margin protected per euro spent, and look at how they perform in the bad case, not only in the median. A response that saves little on average but halves the worst-case loss is often the best insurance.

08How often to repeat it

Run a full stress test once a year and whenever something material changes: a new major supplier, a new market, or a policy announcement such as a tariff proposal. Between full exercises, keep the map up to date and rerun only the scenarios affected by the news. With a model that runs in seconds, rerunning a scenario after a headline is a five-minute job.

09Common mistakes

• Mapping only tier 1. Your German supplier may depend on a single Chinese sub-supplier. Ask your critical suppliers where their critical inputs come from.
• Using averages for lead times. Variability, not the average, drives stock-outs. Record the spread.
• Ignoring cash. Holding six extra weeks of stock may protect delivery but can strain working capital and credit lines.
• One scenario only. The value is in comparing scenarios and responses, not in a single number.
• No owner. Assign one person to keep the map current and to rerun scenarios when the news changes.