Identifying a risk and then writing 'monitor' in the response column is how registers go to die. A response is a decision about who does what, by when, and what 'success' looks like if the risk moves.
The four plays for threats
Avoid: change the plan so the risk cannot happen. Drop the risky integration. Use a proven vendor. Cut the scope that sits on the critical path. Avoidance is often the cheapest option that nobody wants, because it requires saying no.
Transfer: move the impact to someone better placed to hold it — insurance, a fixed-price clause, or a specialist vendor. Transfer is not magic. You usually keep residual risk (the insurer argues, the vendor disputes the clause) and you always keep the relationship cost.
Mitigate: reduce likelihood, impact, or both. A prototype reduces technical uncertainty. A backup resource reduces key-person risk. Mitigation should be a verb with a date, not a slogan. 'Improve communication' is not a mitigation. 'Weekly integration demo with the vendor from 12 May' is.
Accept: choose to live with it, preferably with a trigger. Active acceptance means you watch a metric and have a fallback ready. Passive acceptance means you shrugged. Only use passive acceptance on risks that would not change a decision if they landed.
Opportunities are not a motivational poster
The same four ideas flip for opportunities: exploit, share, enhance, accept. If a regulatory window might open early, exploiting it means having a release train ready. Sharing might mean partnering so you can take the upside you cannot capture alone.
Most teams never log opportunities. They log only the monsters. That trains the organisation to treat uncertainty as purely negative, which is how you miss a cheaper vendor, a reusable component, or a sponsor who would fund a second increment if you asked while the first one is landing well.
Write the opportunity with the same cause–event–effect discipline. 'We might finish early' is not useful. 'If UAT defects stay below the last-release rate, we can bring go-live forward two weeks and avoid the August freeze' is a decision waiting for a trigger.
Pick one on purpose
A risk with three responses and no owner is a committee. Choose the primary strategy, name the residual risk, and fund the action. If mitigation costs more than the EMV of the risk, you are buying comfort, not value. Sometimes that is still rational — reputational risk does not always fit a spreadsheet — but say that out loud.
Fallback plans belong on the high-impact risks. The fallback is what you do if the primary response fails. Dual-source the vendor. Have the rollback script. Know who calls the customer. A fallback invented during the incident is just adrenaline with a job title.
Document the decision. When a risk lands and someone asks why you accepted it, 'we discussed it in March' is not an answer. A one-line rationale in the register is.
Summary & next steps
Threats: avoid, transfer, mitigate, accept. Opportunities: exploit, share, enhance, accept. Pick a primary strategy, name residual risk, and attach an owner and a date. 'Monitor' is not a strategy unless the trigger is explicit.
Next: monitoring and control — how the register stays honest after the workshop glow fades.