Do not begin by asking the delivery team whether they are confident. They have been living inside the programme for months and are usually very motivated to reach the date.
An independent review should test the release argument itself.
Our independent review starts with the risk, finds the evidence and then sees whether the confidence survives. That is rather different from reviewing the slide deck and congratulating everybody on the colour scheme.
Start with what cannot safely fail
Work out which customer journeys, data, payments, integrations, operational processes and technical behaviours would cause serious damage if they failed. Then trace the evidence for those things.
For a critical journey, ‘the team tested it’ is not enough. Which environment? What data? Was it genuinely end to end? Were important failure paths exercised? Did a supplier stub out the awkward bit? Was the test blocked and quietly removed from scope? If that review feeds a go-live decision, make sure it also answers what independent assurance before go-live must tell you.
Performance, security, accessibility, resilience and recovery should be checked where they matter to the service, not added to the review because a generic assurance template has six boxes.
Speak below the reporting layer
Testers, developers, operations staff and technical specialists often describe a more useful reality than the steering pack. That is not because senior people are lying. Information gets compressed, softened and tidied as it travels upward.
A good reviewer follows serious risks back to the evidence rather than assuming the summary is complete.
Look for evidence laundering
Tests marked passed because the defect was accepted. Missing evidence described as ‘covered by process’. A blocked integration quietly removed from the denominator. A supplier reporting 96% execution when the unexecuted 4% contains the only journey that matters.
None of those things necessarily means the release must stop. They do mean the decision-maker should see the real position rather than the polished one.
A focused independent review in practice
For a major release, an independent review can follow a short, disciplined sequence. The output should be a recommendation on the release evidence, not a second project plan.
- Agree the question: Which decision must the review inform, and by when? Identify the three to five failure scenarios that would create the greatest business damage.
- Sample actual evidence: Select a critical journey and trace its test results to the exact build, environment, data and defect history. A dashboard percentage cannot substitute for this.
- Challenge the gaps: Speak separately to test leads, engineers, business users and operations. Ask where the current evidence would fail under production conditions.
- Issue a short written view: State what is credible, what is unproven, the consequence of each serious gap, what could be checked before launch and who owns the decision.
Example finding: ‘Customer-balance migration reconciled to the expected totals in the staging sample. However, reversals for dormant accounts were excluded, and no independent comparison was available for that population. A financial misstatement remains possible. Recommend sampling those records before approval; accountable owner: migration lead.’
Give management that finding while there is still time to act. A report delivered after go-live is a historical document, however splendid its formatting.
Do the review while there is still time to change something
Independent assurance is nearly pointless if it begins two days before launch and every meaningful recommendation would require moving the date. Review early enough to fix serious gaps, then refresh the position close to release.
Also separate evidence that is badly reported from evidence that does not exist. Sometimes teams have done good work and the pack is terrible. Sometimes the pack is beautiful because it is doing all the work.
If you have a major release coming up, get onto us. We can independently review the testing evidence, speak to the people doing the work and give you a short written recommendation before the calendar makes the decision for you.