ERP Integration Failure at Helix Group

I would reject unconditional cancellation and compression. The programme has already failed eight of 14 integration tests, and the failures are concentrated in payroll and inventory-the areas where defects are most visible and least tolerable. Risk workshops are where interface, data-migration, and cutover risks are identified; cancelling them removes assurance at the point of greatest uncertainty. Compressing user-acceptance testing by three weeks reduces detection time and shifts defects into the first pay cycle, which HR has said is unacceptable.

If schedule pressure is real, the steering group should reduce scope, not assurance. Conditions: retain at least the workshops focused on payroll and inventory interfaces; re-baseline UAT with written entry and exit criteria; require a payroll parallel run with full reconciliation; require vendor root-cause analysis and remediation plan for the eight failures; require independent test evidence; and define a go/no-go gate. Non-critical reports or interfaces could be deferred. The decision should record the accepted schedule risk and the residual payroll risk. This is a judgement call: protecting the date by removing risk work increases the chance of a failed go-live, which is more costly than a short delay.

The current 'low likelihood' rating is not credible. Financial due diligence tests solvency, not delivery capability. Eight failed integration tests, a departed client-side analyst, and payroll-critical interfaces indicate higher delivery and concentration risk. The $150k service-credit cap is small against a $6.2m programme and the potential cost of payroll failure, and the absence of a performance bond means limited recovery. The risk should be re-scored with explicit impact on payroll, inventory, and go-live.

Actions before further work: obtain a dated remediation plan with named resources; require evidence of similar payroll integrations; negotiate additional assurance such as escrow, step-in rights, retained payments, or expanded credits; set vendor KPIs and a transition/exit plan; and identify a fallback for the first payroll run. The $400k contingency is only 6.5% of budget and may be insufficient for a parallel run plus manual payroll fallback. Reserve a defined portion for payroll assurance rather than treating contingency as schedule buffer. The steering group should approve the revised assessment and the negotiation mandate.

The appetite should be asymmetric and explicit. For the first pay cycle, the tolerance for missed or incorrect payments should be zero or near-zero; payroll failure affects employees, legal obligations, and reputation. Schedule slippage is undesirable but can be tolerated if it buys assurance. The steering group should state that go-live will not proceed unless the payroll parallel run reconciles within an agreed tolerance (for example, 100% of payee records matched with no critical defects open) and inventory cutover balances are agreed. If those conditions are not met, delay is the default.

The trade-off should be communicated with evidence: a three-week UAT compression may save three weeks but increases the likelihood of payroll disruption; the cost of delay is known and bounded, while the cost of failure is wider. The decision, conditions, and residual risk should be minuted with an owner. This prevents schedule pressure from silently overriding a stated risk appetite.