Decision Rules and Management Judgement

By Dr Riaan Steenberg

A listed company does not fail because nobody had opinions. It fails
when too many decisions depend on opinion at the exact moment when
discipline is needed most.

The original note behind this article was only a sentence: develop a
specialised rule-based engine that predicts a company’s performance or
share price from fundamentals. It is a small idea, but it points to a
larger management discipline. The value is not only in predicting
markets. The value is in forcing management to make its assumptions
explicit.

That is what a rule-based engine does. It turns judgement into
visible logic.

The Problem with Hidden
Judgement

Executives often say they are making evidence-based decisions. In
practice, evidence is frequently filtered through memory, mood,
politics, incentives, and the confidence of the person speaking
last.

This is not a moral failure. It is a design problem.

When the rules behind a decision remain hidden, three things
happen.

First, the organisation cannot learn properly. A result can be
explained away after the fact because nobody can inspect the assumptions
that produced the decision.

Second, accountability becomes vague. People debate personalities
instead of decision quality.

Third, the same decision is made differently by different people, in
different rooms, on different days.

That variability is expensive. In a listed entity it can affect
capital allocation, risk appetite, investor confidence, operating
rhythm, and the credibility of management guidance.

What a Rule-Based
Engine Actually Provides

A rule-based engine is not magic. It is a structured way of
saying:

• if these conditions are true,
• and these thresholds have been crossed,
• and these exceptions do not apply,
• then this action or interpretation should follow.

At enterprise scale this can become technical and automated. At
management scale it can begin as something much simpler: a disciplined
model of decision rules.

For example:

• If revenue is growing but cash conversion is weakening, growth
  quality must be challenged.
• If margins improve while customer concentration increases,
  resilience has not necessarily improved.
• If debt is reduced by delaying necessary maintenance, the balance
  sheet may look better while the business becomes weaker.
• If a division consistently beats budget by underinvesting in
  capability, the current result may be borrowing from future
  performance.

These are not accounting formulas pretending to be strategy. They are
management rules that make interpretation more consistent.

Why Fundamentals Matter

The share price is not the company. It is a market interpretation of
the company under conditions of uncertainty.

Management cannot control that interpretation completely, but it can
understand the fundamentals that tend to shape it: revenue quality,
margin durability, cash generation, capital intensity, debt structure,
customer risk, operating leverage, governance quality, and the
credibility of execution.

A rule-based management model asks a practical question:

What would we have to believe for this business to be worth more,
safer, stronger, or more fragile than it appears?

That question is useful even when the model is imperfect. Especially
then.

The purpose of the model is not to remove judgement. It is to improve
judgement by exposing the conditions under which judgement changes.

A Micro-Model for Management

The useful starting point is not a giant predictive system. It is a
micro-model.

Choose one decision area where inconsistency is costly. Capital
allocation is a good example. So is credit risk, acquisition screening,
pricing discipline, or performance review.

Then write the rules in plain language before turning them into
metrics.

For instance:

• We do not call growth healthy unless cash generation follows
  it.
• We do not reward margin improvement that comes from starving the
  future.
• We do not treat once-off gains as proof of operating
  excellence.
• We do not expand a business model until the unit economics are
  visible.
• We do not accept strategic explanations for problems that are
  actually execution problems.

Only after the rules are written should the organisation decide which
numbers will test them.

This order matters. If the numbers come first, the model can become a
dashboard in search of a purpose. If the rules come first, the numbers
become evidence.

Prediction Is the Wrong
First Promise

The tempting promise is prediction: build a model that forecasts
performance or share price.

That may be possible in narrow ways, but it is the wrong first
promise for management. The first promise should be discipline.

A useful rule-based engine should help leaders:

• separate signal from noise,
• identify deteriorating fundamentals earlier,
• apply the same standard across comparable decisions,
• test whether a result was earned or merely reported,
• and make the organisation’s assumptions easier to challenge.

Prediction can follow. Discipline must come first.

Where the Model Becomes
Powerful

The model becomes powerful when it is reviewed after outcomes are
known.

If the rule said a pattern was dangerous and the business performed
well, the rule must be examined. Was the rule too blunt? Did it miss a
countervailing strength? Was management lucky? Was the timing simply
longer than expected?

If the rule said a position was safe and the business weakened, the
same discipline applies. Which assumption failed? Which metric was too
late? Which exception was overused?

This is where many organisations lose the benefit. They build
dashboards, but they do not build learning loops.

A rule-based engine without review becomes bureaucracy. A rule-based
engine with review becomes organisational memory.

The Management Advantage

The real advantage is not that a rule-based engine always knows the
answer. It does not.

The advantage is that it creates a repeatable conversation about
evidence, assumptions, exceptions, and consequences. It gives managers a
way to say, “This is the logic we are using. This is why we believe it.
This is where it may be wrong.”

That is a better conversation than confidence without structure.

In listed entities, where performance is interpreted constantly by
markets, boards, analysts, employees, and competitors, that discipline
matters. It helps management avoid two common traps: reacting
emotionally to market signals, and ignoring market signals because they
are uncomfortable.

The micro-model sits between those traps. It does not worship the
share price. It does not dismiss it. It treats the market as one signal
among many and tests that signal against fundamentals.

Start Small

The best first version can be modest.

Take ten recurring management decisions. Write the rules that should
guide them. Define the evidence required. Name the exceptions. Review
the outcomes every quarter. Improve the rules.

That is enough to begin.

Over time, the organisation can automate parts of the model, connect
it to financial and operating data, and use it to support more complex
decisions. But the engine starts as a management habit before it becomes
a technology asset.

The habit is simple: make the rule visible.

When the rule is visible, it can be tested. When it can be tested, it
can be improved. When it improves, the organisation becomes less
dependent on heroic judgement and more capable of disciplined
performance.

That is the quiet promise of rule-based engines in management. Not a
machine that replaces leadership, but a structure that makes leadership
more honest.