Out-of-Sequence Progress at Month-End
At the month-end update, the data date is 30 September. The schedule shows 22 activities whose successors have actual starts before the predecessors' actual finishes, and 8 activities with actual finishes but no actual starts. The scheduler's draft uses progress override, producing a forecast completion of 12 December against a 30 November baseline. Three change orders have been approved but not incorporated, and two subcontractors have submitted revised durations that extend their work by 10 days each. The project manager wants to issue 12 December as the forecast while preserving 30 November as the baseline for earned value reporting.
1. How should the 22 out-of-sequence activities and 8 missing actuals be handled, and what are the implications of progress override versus retained logic?
Read the worked answer
Treat the anomalies as data-quality signals first. For each of the 22, check whether the successor's actual start is correct and whether the predecessor's remaining work truly continues. If the successor started early because the predecessor's remaining work is not driving it, retained logic can still model that remaining work. If the predecessor must finish before the successor can proceed, the successor's start is either misreported or the logic is wrong. For the 8 finishes without starts, obtain the missing actual start from field records or correct the status.
Progress override lets uncompleted predecessor work be ignored where a successor has started, which can shorten the forecast and create a false completion date. Retained logic preserves the remaining predecessor work and is the safer default for CPM forecasting. Use progress override only as a documented diagnostic, not as the issued schedule. The scheduler should also check open ends and constraints before accepting 12 December.
2. What should be done about the forecast date, the baseline, and the approved changes before reporting?
Read the worked answer
Before issuing a forecast, incorporate the three approved change orders through change control, with their schedule impact and any added activities or logic. Treat the subcontractors' 10-day extensions as claims to be substantiated, not automatic edits: test them against remaining quantities, calendars, and the driving path. Then update the schedule from the 30 September data date with corrected actuals and retained logic.
Report three things separately: the 30 November baseline, the current approved plan (baseline plus approved changes), and the 12 December forecast with its driving path. Preserving the old baseline for earned value is defensible only if the approved changes are disclosed and their value is tracked; otherwise variances become misleading. If governance allows, re-baseline prospectively for the approved scope and keep the original baseline for historical comparison. The project manager should not present 12 December as a variance against an unchanged 30 November baseline without explaining the change orders.