How do you test a new strategy before you commit to it?

01 / WHYWhy most strategies fail before they begin

Most strategies fail not because they are bad, but because they were built on a single assumption that is no longer true. That assumption was true three years ago, perhaps a year ago, but today it is starting to move, and no one stopped to check.

The assumption can be as simple as "this market grows by two percent a year," as complex as "regulation X will not affect category Y," or as hidden as "consumers in this age group are still looking for the same value." In every case, the assumption was never written down explicitly, never tested explicitly, and therefore went unnoticed when it stopped being true.

A strategy that survives only in today's world is not a strategy. It is a plan.

A strategy review is not a search for approval. It is a systematic search for the fragile assumption, the scenario no one considered, and the signal that already exists on the ground but has not yet made it into the quarterly reports.

02 / DIAGNOSISThe real problem: not a lack of information, but a lack of a system picture

Most decision makers who come to a strategic briefing do not suffer from a lack of information. They suffer from an excess of noise: headlines, reports, analyst forecasts, market reviews, consultant decks, and LinkedIn posts. Each offers one perspective, from one field, on a question that spans several fields at once.

A strategic decision in the world of 2026 does not rest on a single field. It rests on connections between fields:

• Technology changes employment patterns, and employment patterns change demand for real estate.
• Geopolitics changes supply chains; supply chains change insurance pricing.
• Demographics change housing patterns; housing patterns change regional land values.
• Energy changes infrastructure planning; infrastructure planning changes the national investment map.

A strategy that does not test these connections makes its decision on a partial picture. The missing pieces are usually the ones that will determine the outcome.

03 / METHODThe Blue Pumpkin method (S5): five stages for testing a strategy

Blue Pumpkin is the proprietary method of the Strategic Intelligence Lab for reading complex systems. It consists of five sequential stages that translate a global signal into a concrete decision question. Each stage stands on its own, but the value is created by running them in sequence.

1. Signal

   Signal: what happened, and what does everyone see?

   Identify the first point that indicates change: an event, a headline, a data point, or a regulatory decision that has broken the surface. Do not try to interpret it yet; just capture the signal at full sharpness.

2. System

   System: which deep system is moving?

   Ask what mechanism is generating the signal: a shift in the balance of power, a policy change, new demand, or a cumulative demographic pattern. The signal is the foam. The system is the wave.

3. Stakes

   Stakes: who is affected, and what is on the line?

   Map every stakeholder: who pays the price if this is missed, who gains if the right action is taken, and who will block the way even if the analysis is correct. Risk is not measured by a formula, but by the identity of the party that will push back.

4. Scenarios

   Scenarios: where could this go?

   Build 3 to 4 possible scenarios, including one that brings the strategy down. Not a single forecast, but a range you can act within. Each scenario gets explicit assumptions, leading signals, and a strategic response.

5. Strategy

   Strategy: which decisions does this change now?

   Translate the systemic view into action: must-ask questions before the decision, assumptions to challenge, and stop signals that will require a return to review. The result is not a document but a decision ruler.

04 / SCANScanning the strategy against 7 intelligence layers

The Blue Pumpkin model tracks seven layers, spread across three speeds. A strategic decision is tested against each layer separately, and then against the frictions between them.

Slow layers (moving over decades)

• Resources and environment. Energy, water, climate, raw materials, physical geography. The layer that defines the foundation for everything else.
• Geopolitics and power. Supply chains, trade routes, technological sovereignty, the balance of power between major powers.
• Demographics and society. Migration, generational dynamics, family structure, consumption habits, emerging identities.

Intermediate layers (moving over years)

• Economy and capital. Interest rates, markets, currencies, employment, macro policy.
• Real estate and infrastructure. Land, housing, offices, cities, power grids, and logistics.

Fast layers (moving in real time)

• Technology and AI. Enterprise AI adoption, data centers, automation, digital government.
• Perception, narrative, and media. Public opinion, collective story, legitimacy, fear and hope.

For each layer, ask three questions:

1. Does the strategy assume anything about this layer?
2. Is that assumption still true today?
3. Which early signals will indicate that the assumption is ceasing to be true?

05 / SCENARIOSBuilding scenarios: three directions plus a counter-scenario

A scenario is not a forecast. It is a coherent description of a possible world, with explicit assumptions that can be tested in real time. A strategy review requires at least four scenarios:

1. Base Case. Existing trends continue at the expected pace. The strategy under review assumes this from the outset.
2. Acceleration. Positive trends accelerate faster than expected. Can the strategy capture the opportunity?
3. Stagnation. Positive trends stall, but there is no acute crisis. Most strategies do not prepare for this.
4. Counter-thesis. Something in the system moves, deliberately or by chance, and the strategy becomes a costly mistake. This scenario is the real test.

A strategy that survives only in the base case is not a strategy; it is a bet. A strategy that holds in three out of four scenarios is a strategy. A strategy that holds in all of them is usually too generic to be useful.

06 / QUESTIONSMust-ask questions and stop signals

The final stage of the review translates the entire analysis into a practical decision tool. It has two components:

Must-Ask Questions

A list of 5 to 8 questions that must be answered in writing before the decision is made. Each question is framed so that "I don't know" is a legitimate answer, and also an answer that requires action.

Examples of typical must-ask questions:

• What is the one assumption that, if proven wrong, turns the strategy into folly?
• Who is the person who knows the field best and disagrees most with the direction?
• Which signal on the ground already contradicts the strategy today, and how significant is it?
• What is the maximum time frame after which the strategy will be deemed a failure?
• How would the decision change if the time window shrank by 50%?

Stop-Light Signals

A list of signals that, if they appear on the ground, require you to stop and return to review. A stop signal is not "if sales drop." It is specific, measurable, and has a clear threshold:

• "If interest rate X crosses threshold Y, we go back and re-examine the financing assumption."
• "If regulation X is published before date Y, we go back and re-examine the revenue model."
• "If demographic indicator X changes by Z percent, we go back and re-examine the target market."

A strategy without stop signals is a strategy that gets updated only when it is already too late.

07 / CHECKLISTStrategy review checklist

An action summary for testing a new strategy before the decision is made:

1. The decision question has been written down in one sentence.
2. 5 to 10 working assumptions have been written down explicitly, including self-evident ones.
3. Each of the 7 layers has been scanned against the strategy.
4. 3 to 4 scenarios have been built, including a counter-scenario.
5. 5 to 8 must-ask questions have been drafted, and all have an answer (or a documented "I don't know").
6. 3 to 5 stop signals have been defined with measurable thresholds.
7. A follow-up review date has been set (usually 3 to 6 months).
8. A designated "challenger" has been appointed to test the strategy over time.