How to Manage Project Risks
By Theta PM Systems · Updated November 2026
A risk is something that might happen. An issue is something that already did. Teams that skip risk management do not avoid risk — they just discover it later, when it has become an issue with a date attached.
Managing risk is a small, repeatable loop: name what could go wrong, decide how much it would hurt, choose a response, and check again next week. It takes less time than the meeting you hold after something breaks.
Step 1: Build a Risk Register
The register is a table, not a document. Its purpose is to make risks visible and owned, so it should stay short enough that someone reads it every week.
- Description: what could happen, written as a specific event rather than a category.
- Owner: one person who watches it. A risk with three owners has none.
- Likelihood and impact: a simple high/medium/low is enough for most projects.
- Response: what you are doing about it, or an explicit decision to accept it.
- Trigger: the early signal that it is happening, so you notice before the deadline.
Step 2: Score and Prioritise
Scoring exists to decide what to discuss, not to produce a precise number. Multiplying invented probabilities by invented costs mostly creates the illusion of rigour.
| Likelihood | Low impact | Medium impact | High impact |
|---|---|---|---|
| High | Monitor | Mitigate | Act now |
| Medium | Accept | Mitigate | Act now |
| Low | Accept | Monitor | Plan a contingency |
Step 3: Pick a Response
There are only four honest responses to a risk. Naming which one you chose is what turns a worry into a plan.
Avoid
Change the plan so the risk cannot happen — drop the dependency, move the launch, choose a simpler approach.
Mitigate
Reduce likelihood or impact: add a review step, split a large migration, or run a spike before committing.
Transfer
Move the consequence to someone else — a vendor contract, insurance, or an SLA. The work does not disappear, but the exposure does.
Accept
Decide consciously to live with it, and write down why. Acceptance is a valid choice; unrecorded acceptance is just forgetting.
Step 4: Watch Triggers, Not Just the Register
The register describes what might happen. Triggers tell you when it starts. A trigger should be something you can observe in the workspace — a task blocked for more than three days, a dependency with no owner, a milestone with no linked work in front of it.
- Blocked tasks that stay blocked: usually an unowned external dependency.
- A milestone with no tasks before it: the date exists, the work does not.
- One person assigned to most of a critical path: a single point of failure.
- Estimates that keep moving in one direction: scope is expanding without being named.
Step 5: Review on a Cadence
Risk review belongs inside an existing meeting, not a new one. Ten minutes in the weekly delivery review, working down the register, is enough: what changed, what is now an issue, what new trigger appeared.
The register should shrink as well as grow. A risk that has passed its window without occurring should be closed with a note, or it becomes background noise that trains the team to ignore the list.
Frequently Asked Questions
What is the difference between a risk and an issue?+
A risk might happen; an issue already has. Risks are tracked with likelihood and impact and managed with a response. Issues are problems to resolve now and are usually tracked as blocked or escalated work.
What should a risk register contain?+
A specific description, a single owner, likelihood and impact, the chosen response, and a trigger that shows when the risk is materialising. Keep it short enough to review weekly.
How many risks should I track?+
Usually five to fifteen for an active project. If the register is longer than anyone will read, it has become a documentation exercise rather than a decision tool.
Is accepting a risk a valid response?+
Yes, provided the decision is recorded with its reasoning. Accepting a risk deliberately is management; accepting it by not thinking about it is how projects fail quietly.
Who should own a risk?+
One person who is close enough to notice the trigger and senior enough to act. Shared ownership reliably results in nobody watching.
Keep Reading
Start free in minutes
Create a workspace, invite your team, and plan your first project today.