A baseline schedule is ready to submit when it is complete, logical and achievable: all the work is in it, every activity is tied into the network, and the calculated dates reflect how the job will really be built.
The checklist below is the one we use before a baseline goes out, and the one we apply when we review a contractor’s baseline for an owner. It works for Primavera P6 and Microsoft Project alike.
Key points
- Check completeness first: missing scope is the hardest problem to see and the most expensive to find later.
- Open ends, hard constraints and long lags are the three logic problems reviewers look for first.
- Very high float usually means missing logic, not spare time.
- Metrics such as the DCMA 14-point assessment flag symptoms. They don’t tell you whether the plan can be built.
Completeness: is all the work there?
- The whole scope is covered. Not just construction: submittals, procurement, mock-ups, inspections, testing, commissioning, punch list and closeout.
- Contract milestones and phasing match the contract. Every required date, access restriction and turnover sequence appears, with the right dates.
- Owner and third-party work is shown. Design reviews, owner-furnished items, permits, utility connections — with durations the contract supports.
- The structure is usable. A sensible work breakdown structure, with activity codes for area, phase, trade and responsibility, so the schedule can be filtered and understood.
Logic: is every activity tied in?
- No open ends. Every activity except the first has a predecessor, and every activity except the last has a successor. An activity with no successor can slip forever without affecting anything — on paper.
- Relationships make construction sense. Mostly finish-to-start, with start-to-start and finish-to-finish used deliberately. Watch for activities whose start is tied but whose finish drives nothing.
- Lags are short and explained. A 15-day lag for curing or delivery is hidden work. It should be an activity someone can see and status. Negative lags should not appear at all.
- Constraints are limited to the contract. A “must finish on” date overrides the logic and distorts float. Constraints belong on contract milestones, not on ordinary work.
Durations and calendars: are the dates believable?
- Durations are reasonable and measurable. Based on quantities and crew output, not worked back from the end date. Many specifications cap ordinary activities at around 20 working days so progress can be judged each month.
- Calendars are right. Work weeks, holidays and any seasonal shutdowns, with weather allowances handled the way the contract describes.
- Cost and resource loading reconciles. Where required, activity costs add up to the contract value and follow the schedule of values.
Results: does the answer make sense?
- The critical path is continuous and plausible. It should run from the start of the project to completion through work an experienced builder would expect to be critical. A critical path that begins halfway through the job usually starts at a constraint.
- Float looks realistic. A large share of activities with months of float points to missing logic. A schedule where almost everything is critical is just as suspect: it usually means constraints, forced ties or durations fitted to the end date.
- The narrative explains it. Assumptions, calendars, constraints and the critical path are described in a basis-of-schedule narrative.
Where the DCMA 14-point assessment fits
The DCMA 14-point assessment is a set of schedule quality metrics developed by the US Defense Contract Management Agency. It is not a contract requirement on most commercial work, but it is widely used as a health check, and scheduling software can report it automatically.
| Check | Common threshold |
|---|---|
| Missing logic | No more than 5% of activities without a predecessor or successor |
| Leads (negative lags) | None |
| Lags | No more than 5% of relationships |
| Relationship types | At least 90% finish-to-start |
| Hard constraints | No more than 5% of activities |
| High float | No more than 5% of activities with over 44 working days |
| Negative float | None |
| High duration | No more than 5% of activities over 44 working days |
| Invalid dates | No actual dates in the future, no forecast dates in the past |
| Resources | Activities with duration carry resources or cost |
| Missed tasks | No more than 5% finishing later than baseline |
| Critical path test | Adding delay to a critical activity moves completion by the same amount |
| Critical path length index | 0.95 or higher |
| Baseline execution index | 0.95 or higher |
Treat the numbers as prompts, not verdicts. A schedule can pass all fourteen and still be unbuildable — or fail two for perfectly good reasons, such as long procurement activities. The metrics tell you where to look. Judgment tells you what you are looking at.
When should a schedule be baselined?
As early as the contract allows — after award, with subcontractor input, and before significant work has been done. Specifications set the deadline: the State of Idaho’s public works contract, for instance, calls for a CPM schedule within three weeks of award on projects over $1 million.
A baseline submitted months into the job has lost most of its value. Progress has already been made against no agreed plan, and early delays have nothing to be measured against.
This guide is general information about construction scheduling, not legal or contract advice. Your contract governs.