Decision-making series · Part 5 of 6

How to choose a decision model: complexity, impact, reversibility and urgency

21 October 2026 · 6 min read · DecixionIQ

Four characteristics of a decision go a long way to telling you which analytical approach to use and how much governance it needs: what kind of problem it is, how much impact it has, how easily it can be reversed, and how urgent it is.

The four characteristics

Problem. Is it structured and well understood, analytically uncertain (the facts can be worked out but are not yet known), or complex and emerging (cause and effect only become clear in hindsight)?
Impact. Is the effect on the organisation limited, or material?
Reversibility. If it turns out to be wrong, can it be undone easily, or only at great cost?
Urgency. Is there time to plan, or is the decision time-critical?

Where each characteristic points

Characteristic	If the decision is…	Start with
Problem	Structured and well understood	Rational decision-making model, cost-benefit analysis, MCDA
Problem	Analytically uncertain	Decision trees, sensitivity analysis, scenario planning, Bayesian reasoning
Problem	Complex or emerging	Cynefin to confirm the situation, scenario planning, real options, small reversible experiments
Impact	Limited	A lighter process within delegated authority, such as PDCA or cost-benefit analysis
Impact	Material	Decision Quality review, documented independent challenge, verification of approval authority
Reversibility	Easy to reverse	Decide, then monitor and correct (PDCA)
Reversibility	Difficult to reverse	Real options and staged commitment (Stage-Gate), with deeper challenge before the first commitment
Urgency	Planned	The full process for the decision's level
Urgency	Time-critical	OODA or experience-based judgement (recognition-primed), with the record completed and reviewed afterwards

A worked example

Take a decision that is complex and emerging, material, difficult to reverse and planned — entering a new market, say, with a large upfront commitment. The characteristics point to:

Cynefin, to confirm that this really is a complex situation, where experiments and learning beat a single big forecast;
scenario planning, to test the strategy against several plausible futures rather than one; and
real options analysis, to value a staged entry that keeps the right to expand or withdraw.

Because it is material and hard to reverse, it should also go to a higher approval authority, with that authority checked and an independent challenge documented before the decision.

A heuristic, not an algorithm

This is a way of routing a decision to sensible starting tools. It is not a validated scoring method, and it does not determine who has legal authority to approve — that comes from your Memorandum of Incorporation, charters and delegations (see Part 4). Use it to start the conversation about how a decision should be handled, then let the people responsible decide.

Where DecixionIQ fits

When you start a decision in DecixionIQ, it suggests a governance level and an approval authority from what you enter, such as the amount involved, using the thresholds your organisation has set. A higher governance level requires more stages of the decision to be recorded.

Sources and further reading

General information, not legal or professional advice.