Stress-Testing a Plan: Will Your Strategy Survive a Rough Future?

A plan usually gets judged against the future its authors expect. Stress-testing judges it against futures they would rather not think about. The question changes from "Is this a good plan?" to "Under which conditions does this plan stop working, and what would we do then?"

The idea is familiar from banking, where regulators run hypothetical recessions through banks' balance sheets. The same logic works for a strategy, a budget or a big project. This guide explains a simple workshop method from the UK Government Office for Science's Futures Toolkit, with supporting ideas from Shell's scenario practice and RAND's Assumption-Based Planning.

What stress-testing means

The Futures Toolkit describes policy stress-testing, also known as wind tunnelling, as a method for testing policy, strategy or project options against a set of scenarios, trends or future events to see how well they stand up to a range of possible external conditions.

The banking version shows the same shape. The Federal Reserve says its stress test assesses whether banks with $100 billion or more in assets are sufficiently capitalized to absorb losses during stressful conditions while still meeting obligations and continuing to lend. It runs the test annually using a minimum of two different scenarios, and has described it as a forward-looking evaluation of how a hypothetical recession scenario would affect banks' capital ratios. You do not need a regulator's models to borrow the principle: decide in advance what "still standing" means, then run several hard but plausible futures through your plan.

What you need before you start

Three inputs make a stress test useful:

• A clear statement of the plan and its objectives. If people disagree about what the plan is trying to achieve, the test will turn into an argument about goals.
• A set of contrasting futures. These can be full scenarios, or simpler sets of trends and events. Several genuinely different futures are better than one bad one.
• A definition of failure. For a small business this might be running out of cash, losing a key customer group, or missing a deadline that matters.

If you do not yet have scenarios, a quick 2x2 built from two important uncertainties is enough to begin.

The four-step workshop

The Futures Toolkit sets out a four-step session that it says can be done in 1.5 to 2 hours, with up to 16 people in two groups of eight. It notes that consensus becomes difficult in a larger group.

1. Introduce the scenarios, unless the group built them. The toolkit suggests a one-page handout per scenario showing the scenario matrix and the main characteristics in bullet points.
2. Introduce the options being tested. Remind people of the question the work is meant to answer and make sure everyone understands each option and its objectives. Participants can also propose new options that look promising in one or more scenarios.
3. Test the options against the scenarios. Split into groups, one per scenario. Each group imagines that the world is as described and decides, for each option, whether it is robust, redundant or in need of modification, and why.
4. Review the findings. Build a table with the options down the side and the scenarios across the top, record each group's verdicts, and discuss the pattern.

The toolkit adds one rule that is easy to break: groups must work with the logic of their scenario and not change the scenario to make an option work better.

Reading the results

The toolkit's guidance on interpreting the table is straightforward. Options that work well in every scenario could be taken forward. Options that do not work in any scenario should be revised. Options that work in some scenarios but not others can be treated as contingency plans for the right circumstances, or modified to make them more robust.

Shell's Scenarios: An Explorer's Guide describes the same sorting. Options that work well across all scenarios are what it calls robust and will probably form a basis for action. Others are contingent on events, and some can be broken into stages: you follow a strategy up to a certain point and then hold back until the direction of events becomes clear. Staging is often the most practical outcome of a stress test, because it lets you start without betting everything.

Testing the assumptions underneath

Scenarios test a plan from the outside in. It also helps to test from the inside out. RAND's report on Assumption-Based Planning recommends identifying the important assumptions a plan rests on and then defining hedging actions, which it describes as rethinking the organization's plans as though an important assumption had failed.

A simple way to apply this is to list the five assumptions your plan depends on most, then for each one ask: "If this were false next year, what would we wish we had done today?" Any answer that is cheap to do now is a strong candidate for action.

Keeping the test honest

The Futures Toolkit names the main facilitation challenge: avoiding groupthink or defensiveness around favored options. A few habits help:

• Have someone other than the plan's author facilitate.
• Include people from outside the team, especially people with a wider stake in the plan.
• Record verdicts per scenario before discussing them, so early speakers do not anchor the room.
• Treat "needs modification" as a useful result, not a failure.

The toolkit also cautions that a single stress-testing workshop is a rough and ready process, and that a complex plan with contingency plans may need several iterations. In its words, the value comes from thinking through the consequences of an uncertain future and asking the question "What if this future happens?"

A worked example

For example, imagine a hypothetical bakery with two locations that plans to open a third and add wholesale accounts. The owners test the plan against four scenarios built from two uncertainties: whether ingredient costs stay stable or spike, and whether local foot traffic grows or shrinks.

• The third location works in two scenarios and fails badly when costs spike and foot traffic shrinks. Verdict: needs modification.
• Wholesale accounts hold up in all four. Verdict: robust, so start now.
• The modified plan stages the third location: sign a short lease, open with a reduced menu, and commit to a full fit-out only if sales cross a set level within six months.

The owners also write down what they would do if they reached the worst quadrant anyway, which becomes a short contingency plan.

Key takeaways

• Stress-testing, also called wind tunnelling, checks how a plan stands up across several contrasting futures, not just the expected one.
• Define what failure means for your organization before you begin.
• Rate each option in each scenario as robust, needing modification, or redundant, and do not bend the scenario to fit the plan.
• Act on robust options, turn partly robust ones into contingency plans or staged moves, and revise the rest.
• Test the assumptions under the plan too, by asking what you would do if each one failed.
• Expect to iterate; one workshop is a rough first pass.

Sources

• UK Government Office for Science, Futures Toolkit (2024 edition)
• Board of Governors of the Federal Reserve System, Stress Tests
• Shell International, Scenarios: An Explorer's Guide (2008)
• RAND Corporation, Dewar, Builder, Hix and Levin, Assumption-Based Planning: A Planning Tool for Very Uncertain Times, MR-114 (1993)