Strategic Planning Under Global Uncertainty: Robustness Over Forecasting

Uncertainty is a planning condition, not a temporary disturbance

Executives have spent several years waiting for conditions to normalise so that planning can resume. That wait is itself the strategic error. Policy volatility, contested trade rules, security disruption and technology-driven regulatory churn are structural features of the current environment, and planning systems built for a stable baseline degrade in them predictably: forecasts get revised until they are meaningless, capital decisions stall, and the organisation defaults to inertia while describing it as prudence.

Strategic planning under geopolitical uncertainty is not about forecasting better. It is about restructuring how decisions are made so the organisation can commit under a range of futures rather than waiting for one.

Separate the three kinds of uncertainty

Treating all uncertainty the same is why planning cycles bog down. Three types behave differently and require different treatment.

Resolvable uncertainty. The answer exists; the organisation has not gathered it. A regulatory position, a competitor's capacity plan, a customer's renewal intent. The response is research, and it has a deadline.

Bounded uncertainty. The outcome is unknown but the range is describable — a price band, a tariff range, an election with three plausible coalition results. The response is scenario planning with quantified consequences and pre-agreed triggers.

Deep uncertainty. The range itself is unknown. The response is not analysis but design: build optionality, avoid irreversible commitments where possible, and shorten the feedback loop so the organisation learns faster than the environment changes.

Most planning failures come from applying the second response to the first (endless scenarios where a phone call would do) or the first to the third (more research on an unknowable question, producing delay dressed as diligence).

Robustness beats optimisation

A plan optimised for the expected case is fragile by construction. Under uncertainty the better test is not "which option produces the highest return in the base case" but "which option performs acceptably across the plausible range, and which failures can we survive".

Three design principles follow.

Reversibility has value. A leased facility, a contract manufacturer or a joint venture with a clean exit costs more per unit than an owned plant and buys the right to be wrong. Price that option explicitly rather than treating it as inefficiency.

Staging beats sequencing. Break large commitments into tranches with defined go/no-go gates tied to observable indicators, not to calendar dates. The gate is the point of the design; without it, staging is just slower spending.

Diversity has a floor and a cost. Two suppliers in the same industrial cluster are one supplier. Redundancy is only redundancy if the failure modes are genuinely uncorrelated — and beyond a point, additional diversification costs more than the risk it removes. Say where the floor is.

Indicators, thresholds and pre-committed action

The mechanism that converts scenarios into behaviour is a small indicator set with thresholds and pre-agreed actions. It works when it is disciplined:

• Few indicators. Six to ten per material market. Anything larger becomes a monitoring product nobody reads.
• Observable and objective. Airport and port status, export-licence approval times, currency convertibility, insurance pricing, licensing decisions for foreign operators — not sentiment.
• Named thresholds. A level that, when crossed, requires a decision. Without a number it is a newsletter.
• Owners and lead times. Every action has a name attached and a realistic time to execute. An action requiring nine months to implement is not a contingency for a six-week event.

The output cadence is a monthly change log — what moved, what it means, what decision is requested — not a news digest.

Decision rights are the bottleneck

In most global organisations the constraint under uncertainty is not information; it is authority. The country manager can see the change, the regional executive can interpret it, and nobody below the executive committee can act on it — so the organisation observes accurately and responds late.

Fixing this is a governance exercise: define which decisions move to which level at which escalation tier, delegate spending authority in advance for pre-approved contingency actions, and set a review cadence fast enough to match the environment. Firms that shortened their strategy review cycle from annual to quarterly, with a standing exposure item, are consistently faster to reallocate than those that kept a heavier annual process.

What boards should be able to answer

Five questions, answerable in under ten minutes:

1. Which three jurisdictions carry the most value at risk, and how did that change this quarter?
2. What single event would cause the largest unplanned loss, and what is our lead time to respond?
3. Where are we single-sourced or irreversibly committed in a jurisdiction we would not choose today?
4. Which decisions are we deferring because the picture is unresolved — and what is the quantified cost of that delay?
5. Which indicators would trigger a pre-agreed action, and who owns it?

The fourth question is the one most often missing. Deferral is a decision with a cost, and organisations that never price it tend to over-consume it.

The discipline that makes it credible

Calibration. Record judgements with explicit probability language, record the evidence that would falsify them, and score them retrospectively. A planning function that has never graded its own accuracy cannot claim to be improving, and boards are increasingly asking.

Related reading: qualitative geopolitical analysis and scenario planning and our strategy execution framework. For programme design and delivery support see strategic consulting — or book an executive consultation.