The Resource Everyone Agreed to Share
Talbot Manufacturing is running a nine-month shop-floor IoT project. The project manager sits in the PMO with no direct reports. The two integration engineers the plan depends on report to the plant engineering manager, who committed them at 60% to the project at kickoff. For the last three weeks both have been pulled onto unplanned maintenance work; two project milestones have slipped by a combined 11 days. The project manager has raised this informally twice; the engineering manager says maintenance is "revenue-critical" and the project can "wait a week." The RACI approved at kickoff names the two engineers as Responsible for the integration tasks and the engineering manager as Consulted. The sponsor, the VP of Operations, has not been involved since charter sign-off. The project is on the critical path to a customer audit in seven weeks.
1. What should the project manager do now, and why is escalation on its own not sufficient?
Read the worked answer
Escalation is the right move, but only if it arrives carrying the information the sponsor needs to decide.
Do this first:
- Quantify. 11 days lost in three weeks against a seven-week runway. Recalculate the projected completion date from the actual capacity being supplied (roughly 40% rather than 60%), not from the plan. State the date the audit will be missed.
- Document. Raise a formal issue in the issue log, referencing the two informal conversations and their dates. An escalation that begins at "this has been going on for weeks" without a record is a he-said-she-said.
- Confirm the commitment. Retrieve the resource management plan and the kickoff minutes showing the 60% commitment, so the conversation is about a broken agreement rather than about competing priorities.
- Escalate with a specific ask and options. Not "please help" but: restore 60% immediately, backfill maintenance with a contractor, or approve a re-baseline. Give the sponsor a recommended option and the consequence of each.
Why not just escalate: escalation transfers the decision but not the analysis. A sponsor handed a conflict with no numbers will typically split the difference, which leaves the project at 50% of the resource and the PM back here in three weeks.
2. The engineering manager is listed as Consulted while controlling the resource. What does that tell you, and what should change?
Read the worked answer
It tells you the RACI was wrong at kickoff, and that is a project management defect, not just an engineering one.
A person who controls the supply of a Responsible party's time cannot be merely Consulted. Under the current RACI, the engineering manager has all the influence and none of the accountability; the project manager has all the accountability and no authority. That is the structural reason the informal conversations went nowhere - there was no agreed forum in which the commitment could be enforced.
What to change:
- Amend the RACI: the engineering manager becomes Accountable (or at minimum an Approver) for resourcing decisions on the integration work, with the project manager Accountable for delivery. Make the split explicit.
- Update the resource management plan with named individuals, committed percentages, and effective dates - plus what happens when the commitment is breached.
- Get the sponsor to endorse the change in writing. A RACI revised by the project manager alone carries no more weight than the original.
- Log the overallocation as a risk, not only as an issue. The pattern (unplanned maintenance drawing on project staff) will recur; the response should be planned, not improvised.
The judgement: the project manager's first instinct is often to work harder on the relationship. The more useful instinct is to check whether the governance artefacts describe reality. Here they do not, which is why every conversation was a negotiation from scratch.
3. The sponsor confirms the audit date is immovable but will not override the engineering manager's priority. What now?
Read the worked answer
Accept the constraint and re-plan against the capacity that actually exists. Continuing to plan for 60% when 40% is being delivered is the only genuinely wrong option.
Re-plan with real capacity:
- Recalculate durations for integration tasks at 40% and produce the new forecast completion date. Show the sponsor the gap to the audit date explicitly - the immovable date and the real capacity cannot both hold, and that contradiction belongs to the sponsor, not to the project manager.
Then work the options:
- Negotiate a protected window. Three weeks at 100% in exchange for the engineers being genuinely available for maintenance outside it. Often easier to sell than a permanent priority change, and it may fit the audit date.
- Bring in contractors for the integration work: cost, onboarding time, and knowledge-transfer overhead. Realistic only if the audit is more than a few weeks out and the contractors can be productive in time.
- Reduce scope to what the audit actually requires. Distinguish audit-critical scope from phase-two work; deliver the former.
- Re-sequence so that work not dependent on the constrained engineers proceeds in parallel and the constrained path is protected.
In all cases: document the decision, the residual risk, and who accepted it; re-baseline with sponsor approval; and update the risk register. If the audit is missed anyway, the record should show the constraint, the date it was raised, and the decision that was taken - because that is the difference between a managed risk and an unmanaged failure.