Your project has a risk register. It was filled in at the framing stage, it is projected at every committee, and nobody remembers a decision it led to. Project risk management has become a formality, and the cause is neither the tool nor the document template.
It works on one condition: each risk retained leads to a decision, taken by someone with the power to take it, before the date after which acting is no longer useful. Matrices, ratings and colour codes are only means. A useful register therefore contains few lines, worded as precise events, each with an owner, a response written into the plan and a warning signal.
Why a risk register becomes a dead document
A register dies when it has been designed as proof of method, and not as a decision-making instrument. Four mechanisms combine.
- The lines describe concerns, not events. “Risk of delay”, “risk on resources”: nobody can act on such a broad statement, so nobody acts.
- The rating takes the place of treatment. The session is spent debating the move from amber to red. Once the colour is set, the matter seems settled.
- The responses commit to nothing. “Monitor closely”, “raise team awareness”: no task in the schedule, no budget, no date.
- Raising a risk costs the person who raises it. When the author of a warning is called to account, serious risks stop being reported. The register keeps only the harmless lines.
Risk, issue, assumption: three objects not to be confused
A register also deteriorates because it mixes objects that do not call for the same response. The ISO 31000:2018 standard, which provides guidelines for risk management, links risk to the uncertainty affecting objectives; the PMBOK guide speaks, for a project, of an uncertain event. In both cases, a risk has not yet occurred.
| What it is | The question to ask | The treatment | |
|---|---|---|---|
| Risk | An uncertain event, which would affect a project objective if it happened | What do we do beforehand, and what will we do if it happens? | A decided response and a trigger signal |
| Issue | An established fact, which is already hampering the project | Who resolves it, and by what date? | An action plan, then escalation if the level is not sufficient |
| Assumption | What the plan holds to be true without having verified it | Who verifies it, and before which milestone? | A dated verification |
The three are connected. An assumption that nobody can verify before the milestone that depends on it becomes a risk. A risk that materialises becomes an issue: it leaves the register for the action plan.
Project risk management that works, step by step
The sixth edition of PMI’s PMBOK guide organised risk management into seven processes, from planning the approach to monitoring risks. Later editions reorganised the guide around performance domains. This cycle can be reduced to four steps, with, for each, what separates the formal exercise from useful work.
Identify: word a precise event
Everything depends on the sentence. A well-worded risk links a cause, an event and a consequence on an objective: “the vendor has not frozen the content of its release: if it arrives after the acceptance testing freeze, the performance tests cannot be held before the cutover”. This sentence already identifies who can act, and until when.
Look first in the plan’s assumptions, the dependencies, the critical path and suppliers’ commitments, the subject of our article on contractual risk in project governance. Also ask those who do the work: those who report on it see only part of it.
Assess: rank in order to choose
Probability and impact are rated on scales defined in advance, expressed in the project’s terms: timescale, cost, scope, expected benefits. Add a piece of information that is often missing: the date after which it will be too late to act.
The rating measures nothing precisely. It is used to rank, in order to isolate the short list that the committee can examine in one session.
Treat: a strategy, a task, an owner
For a threat, the same edition of the PMBOK guide distinguished five response strategies.
- Avoid: change the plan to remove the cause, for example by removing an item from the scope.
- Mitigate: reduce the probability or impact through anticipatory action, such as a prototype or a cutover rehearsal.
- Transfer: have a third party bear the consequence, through a contract or insurance. Transfer shifts the financial burden, rarely the delay.
- Accept: take no action on the risk itself, with or without a schedule or budget reserve.
- Escalate: entrust the risk to the level that has the power to act, when it exceeds the project’s mandate.
What makes the difference: the response becomes a task in the schedule, with an owner, a date and a cost. Acceptance is a decision taken by whoever has the authority, never an absence of response.
Monitor: keep the list alive between two committees
Each risk retained receives a trigger signal: an observable fact that announces its occurrence and launches the planned response. At a fixed pace, monitoring checks three points: are the response actions being carried out, have the signals appeared, have new risks emerged? It also closes lines: a risk whose deadline has passed leaves the register.
Who decides, when and on what information
A well-kept register has no effect if it is not linked to any decision-making power.
Who. The owner of a risk is the person who has the authority and the means to act on it, not the person who reported it. Written thresholds specify what the project manager handles alone, what falls to the programme director and what only the sponsor can accept, for example a risk that affects the expected benefits or exceeds the reserve.
When. Each risk on the short list carries the date of the last useful decision. The committee’s agenda is built from these dates; the colour of the boxes plays no part in it.
On what information. The committee does not receive a matrix to comment on. It receives a choice: the response options, the cost of each, and what the project faces if nothing is decided.
The project office keeps the register and prepares these decisions, without becoming the owner of the risks. Its design is covered on our page on setting up a PMO. Before a major commitment, an outside view such as a project audit checks that the risks are known to those who decide.
What takes away the register’s decision-making power
- Judging the process by the number of lines. A long register reassures, and drowns the few risks that matter.
- Letting the product of probability and impact decide alone. An unlikely risk that could stop the project can obtain the same score as a frequent inconvenience.
- Entrusting risk management to the PMO alone. Whoever keeps the list does not have the power to decide on the responses.
- Keeping the sponsor’s decisions out of scope. A date imposed without margin or a scope frozen before the study are sources of risk like any other, and must be able to be stated.
The signs of a living process
A living process can be recognised during the project, by five facts.
- The register changes from one committee to the next: lines are closed, reworded, added.
- Each risk on the short list has an owner, a dated action and a trigger signal.
- The issues that arise were in the register before they occurred. The share of surprises says more than the ratings.
- Committee minutes keep a record of decisions taken on risks, and not only of their presentation.
- Teams report risks without being invited to.
Risk management is therefore judged not by the quality of the document, but by the decisions taken in time. If your register triggers nothing, start with the short list, the owners and the dates of the last useful decision. On a programme, some risks arise between projects, and no project register shows them. Dealing with them is a matter for programme management, whose page describes the governance and delegation thresholds that give the register its outlet.