How to Monitor Your Decisions

Monitoring is the deliberate observation of conditions, signals, and thresholds after action begins so a choice can be revised before failure hardens.

Why decisions need active monitoring

For Deciders, the job is not finished when a course of action is chosen. What was assumed at the point of decision must be checked against what is actually occurring. When outcomes diverge from intentions, either the implementation drifted from the intent or the premises no longer hold.

A flawed process needs correcting at the source. An implementation gap needs closing. Degradation needs repair. A shift in context may require the entire decision to be revisited. These are different problems, and no single check catches all of them. That is why the monitoring plan must be designed alongside the decision, not bolted on afterwards.

Why monitoring is the first thing dropped

In nearly fifty years of advisory work, the pattern I have seen most often is this: a decision is made with genuine rigour, and then oversight is treated as someone else's problem. The process model on the wall ends at selection, and the decision leaves the room. The assumptions behind it are never tested again. Months or years later, when the outcome disappoints, no one can reconstruct what was assumed or when those conditions ceased to hold.

The structural reason is that most organisations assign this work to a separate function. The risk register sits in one department. The decision sits in another. The register catalogues hazards in the abstract. It does not connect a specific premise to a specific decision to a specific trigger for revision. When the premise breaks, the register does not notice, because it was never watching in the first place.

The cost of omitting this step is not visible until it is too late. A manufacturer who decided to source from a single supplier assumed that supplier would remain reliable. That premise was never tested. When the supplier failed, the manufacturer had no early warning and no pre-planned response. A decision autopsy conducted after the event would have shown that the monitoring plan was empty. The decision was made. The watching was not.

What monitoring is not

It is not compliance reporting. It is not the quarterly risk review that a board committee receives, filed alongside the financial statements and forgotten. Compliance reporting asks whether someone followed the rules. Monitoring asks whether the world still matches what the Decider relied on. That is a fundamentally different question, and confusing the two is one of the main reasons organisations discover too late that a decision has failed.

The Australian newspaper industry operated for decades on the assumption that classified advertising revenue would remain stable. No one wrote that down. No one assigned it a significance rating. No one specified what would constitute evidence that it was breaking. When online advertising invalidated it entirely, the industry had no trigger, no threshold, and no pre-planned response. Fairfax Media was eventually sold at vastly degraded value. The executives who inherited the original business model had no record of what it rested on, because it had never been articulated, let alone tested.

A report tells you what happened last quarter. Active review tells you whether the conditions your decision depends on still hold. One looks backward. The other looks at the present and compares it, continuously, to what was assumed. When the clock is short, that comparison has to happen in seconds, which is why decision making under pressure demands that the monitoring triggers be built before the pressure arrives.

Monitoring turns a good decision into a good outcome

A decision made with rigour at the point of origin can still produce a poor outcome if the conditions it rested on shift and no one notices. This is the gap that separates people who produce better outcomes from those who merely make sound choices at the moment of choosing. The decision itself is necessary but not sufficient. What follows matters at least as much.

Consider how disruption actually arrives. It does not announce itself as a crisis. It arrives as a slow drift in conditions that were once stable, which is why change management that ignores the original assumptions is just project administration. A key supplier's quality declines over months. A regulatory environment shifts through a sequence of small signals. A technology that underpins your operating model loses relevance by degrees. The Decider who has named the critical assumptions and specified what to watch catches the drift early. The one who treated the decision as settled discovers it only when the outcome has already failed.

Roger Estall and I observed this pattern across mining, aviation, finance, and public health. The organisations that produced consistently good outcomes were not the ones that made the cleverest decisions at the point of origin. They were the ones that watched what they had assumed and acted when the watching told them something had changed. The Universal Decision-Making Method builds this into the final step for precisely that reason. Without this step, a decision is frozen at the moment it was made. With it, the decision stays connected to what is actually happening.

Building the plan into the decision

The plan answers a simple question: what would tell me that this decision needs revisiting? For each significant assumption, the plan specifies what to watch and what triggers a revision. That specification is the difference between a decision that can adapt and one that drifts until it fails.

Roger Estall and I built this into the final step of the method we set out in Deciding, because the moment a decision is finalised without a review plan is the moment it begins to decay. The Decider who has specified triggers and thresholds before leaving the room is already prepared when conditions shift. The one who treated the decision as final is not. Skip that discipline and the eventual post mortem analysis has nothing to compare reality against, and neither does the after action review. As Roger and I wrote: implement and monitor are a single step, not two.

A mining company I advised had been running a tailings dam on the original design parameters for over a decade. The engineers who approved the design had assumed a particular rainfall intensity and a particular rate of sediment accumulation. Neither figure had been revisited. When we applied the method and checked those figures against current data, the rainfall figure was materially wrong: climate patterns had shifted, and the actual intensity exceeded the design tolerance. The company re-engineered the overflow capacity before the wet season. No breach occurred. That outcome was not the product of a better original decision. It was the product of checking a historical decision and asking whether its assumptions still held.

---

Grant Purdy is the co-author, with Roger Estall, of Deciding (2020), and the architect of the Universal Decision-Making Method.