Month 8 on a Lump-Sum Water Plant Automation Contract

Metrics (BAC = $4.20M)

• EV = 0.48 × 4.20 = $2.016M
• PV = 0.57 × 4.20 = $2.394M
• CPI = 2.016 / 2.310 = 0.87
• SPI = 2.016 / 2.394 = 0.84
• EAC = BAC / CPI = 4.20 / 0.87 ≈ $4.81M
• VAC ≈ −$0.61M (−14.5%)

Interpretation with the necessary caveat

On a lump-sum contract, EV must be earned against the contract baseline, not the current working scope. The 450 added points are not yet under contract. If the team has been crediting EV for that work, roughly $150–200k of EV is unbilled scope - and stripping it out takes CPI closer to 0.80. Either way the forecast worsens. Say this explicitly; the headline CPI is optimistic, not pessimistic.

Also note that SPI expressed in cost terms is a weak schedule indicator on a software-and-panels project. The real schedule story is the critical path: contractual FAT in six weeks, panels five weeks late - about one week of float, effectively zero.

Message to the steering committee

1. Forecast at completion is approximately $4.8M, not $4.2M, and is trending worse until the added scope is priced.
2. The 14-month completion date is achievable only through a specific recovery action (split FAT plus panel delivery recovery), not through effort.
3. A change order for the 450 points must be raised and approved now; every week of delay increases the unbilled exposure and weakens the negotiating position.
4. Re-baseline after the change order so the next report is meaningful.

Size it with your own estimate basis. Using a planning rate of ~4 engineering hours per I/O point (hardware configuration, logic, HMI, loop documentation, FAT and site verification, with spares and rework allowance):

• 450 points × 4 h = 1,800 hours
• At a blended $110/h ≈ $198,000, i.e. 4.7% of the $4.2M contract

That is more than double the 2% threshold in the contract, so a written change order is mandatory before the work continues. State the basis you used, because the owner will challenge the rate - the rate is arguable, the 2% threshold is not.

Consequences beyond cost

• Hardware: 450 points at 16 per module ≈ 28 modules, plus spares. Check the installed spare capacity in the panels (typically 20% specified, often less as-built). Analyzers and position feedback often need new racks and cabinet space - and the panels are already five weeks late, so added scope lands directly on the late critical path.
• Schedule: expect several weeks of added panel build and checkout. This is the argument that moves an owner who will not discuss money.
• Database and documentation: I/O list, loop drawings, SCADA database, FAT procedures and as-builts all grow. This is the scope most often silently absorbed.

Process

Issue a formal change request with cost and schedule impact; set a scope freeze date - the date by which the I/O list must be final for the FAT to be held; log every subsequent request; and get the acceptance criteria for the added points agreed up front (which are in FAT-1 versus FAT-2). On a lump-sum contract, absorbing 4.7% of scope silently converts a profitable job into a loss and a closeout dispute.

Split the FAT - and get the contract consequence agreed in writing first.

• FAT-1 (logic and HMI) at the contractual date, using a simulated I/O rack or loop simulator, with the owner witnessing. This exercises the three backwash sequences, alarms, interlocks, and the SCADA database - the bulk of the risk - and keeps the milestone from being held hostage to someone else's panel delay.
• FAT-2 (integrated with real MCCs) after panel delivery, either at the integrator or on site.
• Contractual point: get written agreement that FAT-1 satisfies the milestone for billing and schedule purposes, and that FAT-2 is a defined follow-on activity with its own acceptance criteria. Without that agreement, splitting the FAT protects the calendar but not the commercial position.

In parallel: put the panel delivery under a formal recovery plan with the electrical contractor, and ask whether off-site panel checkout can be done at their works to recover time.

Narrative conflict

Do not let anyone build to two revisions. Specifically:

1. Stop work on the three disputed sequences and raise an RFI/routing item with a response date.
2. Hold a short focused walkdown with the process engineer, the owner's operations lead, and the programmer; resolve each sequence against the process intent, not against who wrote the later revision.
3. Issue a single controlled Rev D with a change log and a named approver. One truth.
4. Configuration management: tag the PLC program revision to the narrative revision and record it in the FAT documentation. If a disputed sequence changes after FAT, it is a change order, not a rework.

The cost of resolving this at FAT is hours; at commissioning it is weeks of site rework and a credibility problem.