RESULT OBTAINED
Canonical method
Structural analysis: ranking the variables before the scenarios
What must be established
A useful analysis keeps its controls visible.
- 01
Delimit the system, its horizon and its perimeter, and list variables that are defined, sourced and free of duplicates.
- 02
Score every cell with a single question: does the variable of the row directly influence the one of the column?
- 03
Check that the ranking holds when the disputed scores are changed.
Decision reference
Make the next choice more explicit.
DECISION SUPPORTED
Which variables really steer the system under study, before scenarios are built?
DELIVERABLE TO KEEP
A short and argued list of driving and relay variables, handed over to the morphological analysis.
NEXT ELEMENT TO EXAMINE
Morphological analysisExplore →Ask a single question of every cell
The quality of the analysis rests on the discipline of the question asked. For each pair, the group asks whether the variable of the row directly influences the one of the column, and with what intensity: none, weak, moderate or strong, scored from 0 to 3. Two words carry the weight. Directly rules out the influences that travel through a third variable, since the computation of indirect influences will recover them later; including them here would amount to counting them twice. Influence is neither correlation nor proven causality: it is a judgement, and it is written as such, with its justification and its level of confidence. The list of variables settles the rest: two overlapping variables inflate their shared driving power artificially, a variable drawn too wide absorbs the others, a missing variable will never reappear. Each one therefore receives a short definition, a perimeter and a source. The diagonal stays empty: a variable does not influence itself.
- Direct influence
- Score from 0 to 3
- Justification and confidence
- Definitions without overlap
Read the map without turning it into a verdict
The sum of a row gives the driving power of the variable, that is, what it sets in motion; the sum of a column gives its dependence, what sets it in motion. Placed on a map, with dependence on the horizontal axis and driving power on the vertical one, the variables fall into four zones: driving, weakly dependent and strongly influential; relay, at once influential and influenced, therefore unstable; dependent, which undergo; autonomous, barely connected to the rest. The thresholds that separate these zones are a convention, usually the mean of the scores in the matrix: moving them moves the borders, and a variable close to a threshold may change zone although nothing has changed in the system. The map therefore serves to debate, not to settle. A variable is neither retained nor discarded on its position alone: it is so by an argument, and the exclusion of a driving variable is justified in writing. Relay variables deserve particular attention: they are the ones that propagate shocks.
- Driving power and dependence
- Four zones
- Conventional threshold
- Argued exclusion
What indirect influences reveal
A variable may look unobtrusive and weigh heavily by rebound: it influences a variable that influences many others. These paths are recovered by multiplying the matrix by itself, twice, three times, then by observing the ranking obtained at each power. In practice the ranking stabilises quickly; the work stops once it no longer changes from one power to the next. What matters is not the final ranking, but its gap with the direct one: a variable that gains several places acts mainly through its relays, and it would be imprudent to treat it as secondary; a variable that loses places owes its apparent rank to an influence without continuation. This test completes the one on stability: the most debated scores are altered, then the retained list is checked for change. If it changes, the list is not ready, and it is the debate that must be resumed, not the calculation.
- Indirect paths
- Stop at stability
- Gap in rank
- Test of disputed scores
DECISION ASSET
Matrix of direct influences
Every pair of variables receives a direct influence rating, then the row and column sums place the variable on the plane.
Variables
- Input
- Definition and source
- Control
- Duplicate or overlap?
- Output
- Settled list
Rating
- Input
- Influence from 0 to 3
- Control
- Direct influence?
- Output
- Completed matrix
Sums
- Input
- Rows and columns
- Control
- Driving or dependent?
- Output
- Position on the plane
Indirect
- Input
- Powers of the matrix
- Control
- Does the rank change?
- Output
- Rank shift
- CICDE, Operational foresight, why and how, 2022 (opens in a new tab)Public document describing the matrix from 0 to 3, the row and column sums and the indirect influences.
- IWEPS, What are the different types of variables? (opens in a new tab)Institutional definition of driving, relay, dependent and autonomous variables.
MOVE FROM ANALYSIS TO YOUR DECISION
Produce a testable decision mandate
MOVE FROM ANALYSIS TO YOUR DECISION
Produce a testable decision mandate
RETURN TO THE DECISION SYSTEM
See the canonical definition of marketing strategy→