methodEnglish

SPE Structure and Cross-Impact

How to build the variable structure and encode causal influence through cross-impact patterns.

EnglishEspaƱolFranƧais

Documentation follows the current system language. Missing translations fall back to the page's canonical language.

SPE Structure and Cross-Impact

The SPE structure is a causal model. Its purpose is not to list every possible factor, but to represent the variables that explain how strategic choices change outcomes.

SPE causal layers

Strategies enter through Drivers; External Agents define context; Processes mediate effects; Outcomes terminate the chain.

For a first model, use three operational layers:

  1. Drivers: the variables strategy can directly affect.
  2. Processes: the mechanisms that transform Driver effects.
  3. Outcomes: the final metrics used for comparison.

Use Level 0 External Agents when the model needs scenario context. The usual beginner sequence is Level 1 Drivers, Level 2 Processes, and Level 3 Outcomes. This order is the recommended point of departure for new modelers. It is not a hard SPE rule. More advanced models may use additional levels or place Drivers, Processes, Outcomes, or External Agents in different levels when the causal logic requires it.

Add External Agents when the model needs scenario context such as market demand, regulation, economic conditions, climate uncertainty, supply volatility, or political environment.

Variable quality test

Each variable should pass these tests:

  • it has one clear role
  • it is specific enough to assess
  • it can be interpreted on the NN, N, O, P, PP state scale
  • it is not a duplicate of another variable
  • it has a plausible path to one or more Outcomes
  • its direction of influence can be discussed by experts

Cross-impact relationships

A cross-impact relationship answers this question:

If the source variable changes state, how does that influence the target variable?

Use the seven impact patterns to encode direction and strength:

  • Sig+: significant same-direction influence
  • Mod+: moderate same-direction influence
  • Sli+: slight same-direction influence
  • No: no material influence
  • Sli-: slight opposite-direction influence
  • Mod-: moderate opposite-direction influence
  • Sig-: significant opposite-direction influence
Seven default impact strengths

The sign gives direction and the magnitude gives strength. The No pattern has zero effect.

Modeling discipline

Do not connect everything to everything. A dense matrix can look sophisticated but often hides weak reasoning. Prefer a model where each relationship has a concise rationale and where missing links are intentional.

Review questions

  • Which variables are direct strategic entry points?
  • Which variables only mediate effects?
  • Which variables are final outcome metrics?
  • Are any External Agents being treated incorrectly as strategy-controlled?
  • Does every Driver connect to at least one downstream Process or Outcome?
  • Does every Outcome have at least one plausible upstream path?

References

Strategic Decisions Method Comparing Risks, Performance Outcomes, and Scenarios

Internal SPE technical source describing the Strategic Performance Engine framework, variable structure, cross-impact logic, strategies, scenarios, Monte Carlo simulation, sensitivity analysis, and outcome comparison.

SPE Core Variable Definitions

Albarena internal canonical note defining External Agents, Drivers, Processes, and Outcomes for SPE model construction.

Project Performance Modeling: A Methodology for Evaluating Project Execution Strategies

Alarcon-Cardenas, L. F., and Ashley, D. B. Construction Industry Institute Source Document 80, 1992.

Modeling Project Performance for Decision Making

Alarcon, L. F., and Ashley, D. B. Journal of Construction Engineering and Management, ASCE, 1996.

Computer Aided Strategic Planning in Construction Firms

Alarcon, L. F., and Bastias, A. Paper describing the computer-supported strategic planning implementation of the performance modeling methodology.

Un ambiente integrado para la modelacion de decisiones estrategicas

Bastias Largo, A. G. Master thesis, Pontificia Universidad Catolica de Chile, 1998.

Performance Modeling for Contractor Selection

Alarcon, L. F., and Mourgues, C. Journal of Management in Engineering, ASCE, 2002.