However complex the project, I found using the simple 9 box (likelihood versus impact} risk matrix more than adequate for use and presentation
Also making Likelihood and Impact scores project specific (not generic) was useful eg:
- Score Likelihood (1, 2 or 3) against how many project objectives the risk impacts (eg none, one, more than one)
- Score Impact (1, 2 or 3) using pre-determined metrics (eg for cost £50k, £100k, £500k). Could also be weeks, complexity of technology/fabrication, threat of unsafe events etc
I developed an initial matrix (usually by brainstorm) during front-end engineering and estimating. If necessary it could be used to back up project contingencies in the cost estimate
At sanction I would split it into two for the rest of the project:
- One for the EPC phase (managed with contractors)
- One for commissioning, operation and maintenance (managed with the end-user)