Test Team Pulled Mid-Release
MedDevice Innovations is eight weeks from releasing a software update for a regulated infusion pump. The six-person QA team has lost three testers; their functional manager reassigned them to a higher-margin product and says the move is 'temporary.' Thirty critical defects remain open, automated test coverage is 40%, and the last release had a 12% defect-escape rate against a 2% target. Rework already consumes 18% of project effort, and testers are working 50-hour weeks. The project manager has no direct authority over the testers. The sponsor wants to hold the release date because a hospital customer has threatened to switch suppliers. The quality plan requires full regression testing and a signed test summary before release. The functional manager reports to the same director as the project manager and has refused two informal requests to return the testers.
1. Categorize the project's quality costs and interpret the 12% escape rate and 18% rework. What change in quality strategy do these numbers justify?
Read the worked answer
Quality cost categories: prevention (training, test automation, design reviews), appraisal (testing, inspections), internal failure (rework and defect fixes before release), and external failure (escapes, recalls, regulatory action). The 18% rework is internal failure; the 12% escape rate is external failure and is far more expensive per defect because it reaches a hospital. A 12% escape rate against a 2% target signals that the project is buying quality through appraisal and paying for failure, not preventing defects. The response should not be simply 'test more.' Add prevention: root-cause analysis of the escaped defects, requirements and code reviews, automated regression coverage rising from 40%, and defect injection analysis by phase. Track cost of quality as a ratio-failure cost should fall as prevention rises. For the release decision, use defect density and severity trends, not just test hours; 30 open critical defects is a quality gate issue.
2. How should the project manager resolve the resource conflict with the functional manager? What should happen if negotiation fails?
Read the worked answer
Use interest-based negotiation, not escalation first. The functional manager's interest is the higher-margin product; the project manager's interest is the regulated release. Quantify the risk: three testers are 50% of the QA team, regression coverage is 40%, and 30 critical defects remain; losing them threatens the release date and regulatory evidence. Propose options: return two testers for four weeks and one for two weeks; trade a non-critical test activity; use the vendor's test engineers; or reduce release scope. Put the agreement in the resource calendar and a RACI so commitments are visible. If the functional manager refuses, escalate to the shared director with a one-page impact assessment and the options already offered-not a complaint. Also review the matrix organization's resource-planning process: the reassignment should have gone through capacity planning. Protect the test team from 50-hour weeks; fatigue increases escapes.
3. The sponsor proposes skipping full regression testing to protect the release date. How should the project manager respond?
Read the worked answer
Decline the proposal to skip regression testing, and do not make the decision alone. Explain the consequences in business terms: the product is a regulated infusion pump; skipping regression removes evidence required by the quality plan and may breach regulatory submission commitments. A defect that reaches a hospital is an external failure cost, possible recall, and reputational damage far larger than a late launch. Offer alternatives: (1) reduce scope to a subset of features with full regression on that subset; (2) phase the release; (3) add qualified test resources and automate the highest-risk regression cases; (4) negotiate a short, contractual delay with the hospital customer. If the sponsor still wants to accept the risk, require documented risk acceptance at the appropriate authority, with the quality and regulatory functions involved, and record it in the risk register. The project manager's duty is to make the risk visible, not to sign off on a quality gate that the plan prohibits.