The Shared Tester
In a matrix organisation you manage a customer-billing replacement with a committed go-live in seven weeks. System testing is on the critical path: 200 of 620 test cases remain, and the only performance tester on the team is shared with a production-support group. On Monday the functional manager reassigns that tester to a severity-one production incident for three weeks, describing the move as 'temporary and non-negotiable'. No backup tester is trained on the billing test harness. The test lead reports that 34 of the executed cases already show latency above the 1.5-second threshold at peak load, and that defect fixes are taking longer than planned. Your schedule reserve is 12 days; the contract carries penalties of USD 25,000 per week of delay. The functional manager reports to the same programme director as you.
1. Quantify the impact of the reassignment and set out the recovery options with their costs and limits.
Read the worked answer
Size the loss before negotiating. The test lead's throughput is the controlling number. At roughly 25 cases per week, 200 remaining cases need eight weeks of tester time. You have seven weeks of calendar, and the tester is unavailable for three of them. So the reassignment does not shave the plan - it removes about 75 cases of capacity and leaves roughly 100 cases with nobody to execute them.
The quality data changes the nature of the work. 34 of about 420 executed cases (roughly 8%) already breach the 1.5-second peak threshold. That means the remaining performance work is not 'finish running scripts'; it is 'find and fix latency defects', and fixes have a longer tail than planned. Options:
- Do nothing. Slip of roughly 2.5-3 weeks; the 12-day reserve cannot cover it; exposure of USD 25,000 per week and reputational damage on a committed date.
- Escalate to the programme director. Necessary, but escalation is a decision route, not a recovery plan - it allocates the problem, it does not execute test cases.
- Negotiate a partial return. For example three days a week to billing, two to the incident. Cheapest and fastest, and the incident still gets senior test capability for the hardest two days.
- Train a backup on the harness. One to two weeks of shadowing; no help in week one, but it removes single-point dependency for the remainder.
- Buy a contract performance specialist. Ramp-up and access approvals cost you several days, but it recovers throughput in weeks two and three.
- Re-sequence. Run functional, security and interface tests now, hold performance testing until the tester returns, then compress with bounded overtime.
The point to hold onto: the constraint is not effort, it is a scarce skill on the critical path.
2. What course of action do you recommend to the programme director, and how do you frame it given that the functional manager reports to the same person?
Read the worked answer
Recommend a bundle, not a complaint. A single request to 'give the tester back' will lose to a severity-one incident; a package that protects both outcomes is harder to refuse.
The package:
- Negotiated split, in writing. Three days a week to billing for three weeks, two days to the incident. The incident keeps the senior capability where it is genuinely needed, and the critical path keeps 60% of its throughput - roughly 45 cases recovered of the 75 lost.
- Backup trained in parallel. Start shadowing immediately, even with the split, so the dependency does not recur in the next release.
- Contract specialist for weeks two and three, funded from the contingency, to cover the residual gap.
- Re-sequence functional and interface testing into the three weeks so the calendar is used, not merely survived.
- Spend the 12-day reserve deliberately, on the performance test window at the end, not spread thinly across the plan.
How to frame it to the programme director. Speak in shared-programme terms, not departmental ones: the incident has a service-level exposure the functional manager is accountable for, and the billing programme has a USD 25,000-per-week contractual exposure plus the go-live commitment. Both report to the same person, so the director is the only actor who can price the trade-off across both. Present two numbers - days of billing critical path lost and service risk of a two-day-a-week incident response - and let the director choose with full information.
What not to do. Escalate emotionally, refuse the reassignment you cannot refuse, or quietly ask the team to work weekends. The first damages the relationship, the second is empty, and the third hides a resourcing decision inside individual goodwill and burns the people you will need in week seven.