Skip to main content

Strategic Foresight Map

Problem: A trend report can describe change without improving a decision. A vivid vision can inspire people while hiding the assumptions needed to reach it.

Question: Which future are we preparing to create, what else could plausibly happen, and what should we do now?

Decision: Choose one reversible strategic option and the evidence that would justify, redirect, or stop it.

This method produces a horizon-bound vision, a small scenario set, tested assumptions, a backcast, strategic options, triggers, and one proof-producing next action.

Inputs​

  • One consequential decision and its beneficiary.
  • A named horizon: near enough to affect present choices, far enough for the system to change.
  • Current evidence about actors, flows, constraints, and dependencies.
  • Signals from more than one relevant domain, with source and confidence.
  • A human decision owner who can approve commitments and setpoint changes.

Stop if the beneficiary, decision, or human owner is missing. Foresight without a decision becomes observation; foresight without authority becomes theatre.

Steps​

1. Frame the decision​

Write the decision, beneficiary, horizon, stakes, and system boundary in one paragraph. Separate what is controlled, influenced, and exogenous.

Output: a frame narrow enough to guide action and broad enough to include the real system.

2. Map the present system​

List the actors, incentives, rules, bottlenecks, and flows of value, data, talent, trust, and authority. Mark hidden dependencies and feedback loops.

Output: a current-state map with evidence beside each consequential claim.

3. Scan and rank forces​

Collect structural forces, trends, demographic shifts, weak signals, and plausible shocks. Record the source date, confidence, and why each signal could change the framed decision. Rank forces by impact and uncertainty; do not rank by novelty.

Output: a short force register that separates observed change from inference.

4. Build plausible scenarios​

Combine the highest-impact uncertainties into three or four distinct futures. Include a baseline, an upside, a stress case, and a discontinuity only when the system is exposed to one. Describe how forces interact and which decision changes in each scenario.

Output: a scenario set that changes choices, not four versions of the same forecast.

5. Expose assumptions​

State what the preferred future and current strategy require to remain true. For each assumption, name a falsifier, leading signal, and consequence of being wrong. Do not convert a preference into a prediction.

Output: an assumption ledger with evidence and confidence bounds.

6. Backcast from the preferred future​

Describe the preferred reality at the named horizon. Work backward through capabilities, partnerships, data, rules, behaviours, and milestones that must exist. Return to the present and identify the earliest dependency that can be tested.

Output: a milestone path from future conditions to present capability gaps.

7. Create options and triggers​

Define options that work across several scenarios. Keep commitments reversible while uncertainty is high. For each option, name the signal, threshold, or review date that activates, expands, redirects, or stops it.

Output: an option table with triggers, owner, cost boundary, and kill signal.

8. Commit to the next proof​

Choose the smallest action that tests the most consequential assumption. Name the owner, observation window, gauge, and decision the evidence will reopen.

Output: one action that can leave observable evidence before the larger commitment.

Checks​

  • Dream, Vision, and Prediction remain distinct: the Dream names what is worth creating; Vision makes it visible at a horizon; Predictions state what may happen with calibrated confidence.
  • Every scenario changes at least one option, trigger, or assumption.
  • Every important assumption has a falsifier or is labelled unknown.
  • The preferred future does not erase people affected by the decision or conceal who retains authority.
  • The backcast identifies capabilities and evidence, not only dates and aspirations.
  • The next action is voluntary, proportional, reversible, and tied to a named proof signal.
  • A review date or event is set before execution begins.

Worked transfer​

New AI-native Venture​

Decision: Should a founder test an agent-managed quoting service for small trade firms?

At an 18-month horizon, one operator supervises quoting agents through approval gates and traceable margins. The critical assumption is that firms will share enough job data to improve quote quality while retaining final authority. The first action is five owner interviews and one approval-gated prototype. First proof is one owner choosing the generated quote over their current draft and explaining why.

Existing SME Journey​

Decision: Which dispatch decision should a service business augment first?

At a 12-month horizon, dispatchers see ranked options but approve every reassignment. The critical assumption is that job, travel, and capability data are reliable enough to support ranking. The first action is one week of shadowing. First proof is a measured delay cause and one suggestion a dispatcher says they would use.

The examples transfer the method only if each team replaces the frame, evidence, assumptions, and proof with its own reality.

Failure Modes​

  • Single-future certainty: treating the preferred future as the forecast. Add plausible alternatives and confidence bounds.
  • Trend collection: recording interesting change with no affected decision. Remove any force that changes no assumption, option, or trigger.
  • Scenario theatre: changing the story while every plan stays the same. Rebuild scenarios around uncertainties that alter choices.
  • Backcast fiction: setting milestones without present evidence or capability owners. Test the earliest dependency first.
  • Automation by default: using the vision to remove human authority without evidence, disclosure, or consent. Restore the decision right before proceeding.
  • Permanent watchlist: scanning forever while avoiding action. Freeze one review window and run the smallest useful test.

Proof of done​

The map is complete when a cold reader can identify the decision, beneficiary, horizon, present evidence, scenarios, critical assumptions, preferred future, options, triggers, human authority, and next proof without verbal explanation.

The method has not proved the preferred future. It has worked when the first observation changes confidence, an option, a trigger, or the next question.

Changes my mind: If teams using this map make no better distinction between aspiration, probability, and evidence—or no faster correction after signals change—the method is adding ceremony rather than judgment.

Retrieval​

Use this map before a consequential Venture allocation, an SME operating-model change, or any strategy decision whose value depends on how the surrounding system may change.

Version delta: This first version separates Vision from Prediction and connects a preferred horizon to options, triggers, and observable first proof.

Context​

  • Dreams — choose the preferred reality, describe its horizon, prepare its gaps, and turn the picture into evidence.
  • Predictions — calibrate probabilistic claims without mistaking them for the preferred future.
  • Opportunity Scanner — inspect industry-force gaps when the strategic opportunity is still unclear.
  • Strategy — connect foresight to positioning, priorities, and coordinated action.

Questions​

Next question: Which observation would make us change the option before sunk cost makes the decision for us?