Multi-Dimensional Risk Appetite: One Statement Is Not Enough

Illustration for: Multi-Dimensional Risk Appetite

“The organisation has a low appetite for risk.” I have read that sentence, or its close cousins, in more board papers than I can count, and it has never once helped anyone make a decision. It is not an appetite statement; it is a mood. Real organisations hold several different appetites at the same time, for different kinds of risk, and the whole value of the exercise lives in saying so explicitly.

Why the single statement fails

A business exists to take risk. Every product launch, hire, price change and market entry is a risk taken on purpose in pursuit of return. An organisation that genuinely had a low appetite for all risk would never ship anything.

So a blanket “low appetite” statement is false the moment it is written, and everyone in the building knows it. That has two corrosive effects. First, the statement gets ignored, because it visibly does not describe how the organisation actually behaves. Second, and worse, it gives decision-makers no help with the question appetite exists to answer: which risks should we be taking more of, and which should we be taking less of? A single number cannot answer a question about balance.

Appetite as a profile, not a number

The fix is to state appetite per dimension: a small set of risk categories, each with its own appetite level, its own rationale, and its own measures. The UK Government’s Orange Book uses a five-point scale that works well in practice: averse, minimal, cautious, open, and eager. The labels matter less than the discipline of choosing a different position on the scale for different dimensions.

A typical set of dimensions for a mid-size organisation:

Dot chart showing seven risk dimensions positioned at different points on the five-level appetite scale from averse to eager

An appetite profile: the value is in the spread, not in any single position.

Six to eight dimensions is the practical ceiling. Beyond that the profile stops being memorable, and appetite that nobody can remember is appetite that nobody applies.

Here is the test of honesty: at least one dimension should sit at open or eager. If every row says averse or minimal, you have written the blanket statement again with extra rows. The profile only means something because it says “more risk here, less risk there” in the same breath.

Making each dimension usable

An appetite level on its own is still only a word. Each dimension earns its keep when three things hang off it:

• A rationale the board actually debated. One sentence on why this dimension sits where it does. The debate is the point; the sentence is the receipt.
• Two or three tolerances that make it measurable. “Minimal appetite for cyber risk” becomes: no unsupported systems holding client data, critical patches inside 14 days, no supplier with data access unassessed beyond 12 months. Each tolerance carries a threshold and a red line, and each threshold feeds a reportable indicator.
• A named owner. Someone who is accountable for reporting position against appetite for that dimension, and for escalating when the organisation drifts outside it.

This is the same appetite-to-tolerance-to-threshold cascade that makes any appetite statement real, run once per dimension instead of once in total. The work scales linearly; the usefulness scales much faster.

The trade-offs are the point

The most valuable conversations a multi-dimensional appetite produces are the collisions. Product wants to ship weekly; operational continuity wants change windows and rollback plans. Growth wants a new market entered this quarter; regulatory wants the licensing position confirmed first. These tensions exist whether or not you write them down. The appetite profile forces them into the open, where they can be decided deliberately instead of won by whoever shouts last.

Two practical rules help. When a decision sits inside a single dimension’s appetite, the risk owner decides and records it. When a decision crosses two dimensions with different appetites, it escalates, and the more averse dimension sets the default until someone senior consciously overrides it. That single routing rule resolves most of the arguments before they start.

Reporting against the profile

Appetite reporting fails when it collapses back into one number, a single organisation-wide RAG that hides exactly the differences the profile was built to expose. Report each dimension separately: current position against appetite, trend, and which tolerances are breached or close. A dimension sitting outside appetite is not automatically a crisis; sometimes the honest response is to fix the exposure, and sometimes it is to admit the appetite was set wrong and revise it. Both are legitimate outcomes, and the second one is not a failure. An appetite that never changes with strategy is a poster, not a tool.

Review the profile annually at most, and always after a strategy shift, an acquisition, or a serious incident. The incident review question is underused: did we breach appetite, or did the appetite fail to anticipate this? The answers drive different fixes.

Putting it into practice

Do not start from a blank framework. Start from the last quarter’s real decisions: the three or four judgement calls the executive actually made about risk, whatever they were. Draft a profile that would have guided those decisions, with five or six dimensions and an honest spread across the scale. Then test it backwards: for each of those past decisions, would this profile have changed the call, sped it up, or pushed it to the right person? If the answer is no three times running, the dimensions are wrong or the statements are too vague. When it starts changing answers, cascade the tolerances, wire up the indicators, and put the profile in front of the board once a quarter. One page, one row per dimension, no single number anywhere on it.