The DCMA 14-Point Assessment, Explained

Almost every owner, prime contractor, and federal agency now judges a construction schedule by the same fourteen tests. Here is what each one measures — and what a failure is actually telling you.

The DCMA 14-Point Assessment came out of the Defense Contract Management Agency as a way to sanity-check a contractor's schedule without re-planning it by hand. Be clear about what it does and doesn't do: it doesn't tell you whether the plan is smart. It tells you whether the schedule is built honestly enough that the dates are worth believing. Those are different questions, and conflating them is where a lot of arguments start.

Fourteen checks. Most carry a pass/fail threshold — miss it and that check fails. Below is what each is looking for, in the order DCMA numbers them, with the standard default threshold. (Owners often tune thresholds to their spec; good software lets you match whatever the contract says.)

How to read the checks

They come in three flavors. Percentage checks tolerate a small share of imperfect activities — usually 5%. Zero-tolerance checks fail on a single instance. And two index metrics — CPLI and BEI — score realism and progress rather than counting defects. Keep the distinction in mind: a 4% reading on a percentage check is a pass; a single negative lag is not.

Checks 1–4: Logic & relationships

Whether the network actually behaves like a network.

• 1 — Missing Logic. Threshold: ≤ 5%. Every incomplete activity should have both a predecessor and a successor. Danglers float free of the network and quietly corrupt the critical path.
• 2 — Leads (negative lags). Zero tolerance. A negative lag lets a successor start before its predecessor finishes — the easiest way to shave a date without re-planning the work. Full breakdown here.
• 3 — Lags. Threshold: ≤ 5% of relationships. Positive lags bury work (or padding) inside a relationship line where no one can see or status it. A few are fine; a schedule full of them is hiding activities.
• 4 — Relationship Types. Threshold: ≥ 90% Finish-to-Start. FS logic is how construction actually sequences. A pile of Start-to-Start and Finish-to-Finish links usually means missing detail dressed up as overlap.

Checks 5–7: Constraints & float

Whether the dates are driven by logic or by wishful thinking.

• 5 — Hard Constraints. Threshold: ≤ 5%. "Must Finish On" and mandatory constraints override logic and freeze dates in place. They're wish dates wearing a plan's clothing — and where reality goes to hide.
• 6 — High Float. Threshold: ≤ 5% over 44 working days. Big float numbers feel like comfort but usually signal missing logic, not genuine slack. An activity with 80 days of float is often one nobody finished wiring up.
• 7 — Negative Float. Zero tolerance. The schedule telling you, in red, that the contract finish is already unachievable as logicked. Don't edit it until the red goes away. Full breakdown here.

Checks 8–10: Durations, dates & resources

Whether activities are sized and statused so the schedule can be managed.

• 8 — High Duration. Threshold: ≤ 5% over 44 working days. Long activities hide detail, resist honest status, and smear progress across months. Break them down or you can't tell a half-done bar from a stuck one.
• 9 — Invalid Dates. Zero tolerance. No forecast dates in the past and no actual dates in the future, relative to the data date. Violations mean the schedule wasn't statused cleanly — and every date downstream is suspect.
• 10 — Resources. Threshold: all worked activities loaded. On a resource- or cost-loaded schedule, every activity with work should carry resources or cost. Activities missing them never show earned value — a blind spot in your reporting.

Checks 11–14: Execution health

Whether the schedule is keeping pace with its own baseline.

• 11 — Missed Tasks. Threshold: ≤ 5%. Activities that finished later than the baseline said they would. The earliest objective warning that the plan is slipping.
• 12 — Critical Path Test. Pass/fail. An integrity test: inject a large delay into a critical activity and the project finish should move by the same amount. If it doesn't, your critical path is broken — open ends, a stray constraint, or missing logic.
• 13 — Critical Path Length Index (CPLI). Target ≥ 0.95. (Critical path length + total float to the finish) ÷ critical path length. Below 0.95 means you're counting on burning float to make the date.
• 14 — Baseline Execution Index (BEI). Target ≥ 0.95. Tasks actually completed ÷ tasks that should have been complete by now per baseline. Under 0.95 means you're finishing work slower than you planned to.

What a passing score doesn't tell you

Passing all fourteen doesn't mean the schedule is good. It means it's built cleanly enough to trust the mechanics. A schedule can ace the 14-Point and still sequence work in a way that will never happen in the field — the checks catch dishonesty and sloppiness, not bad strategy. But the reverse is reliably true: a schedule that fails the 14-Point almost always has bigger problems hiding behind the findings. That's why reviewers run it first.

Why it became the standard

Owners and reviewers needed an objective, repeatable screen they could apply to any contractor's file without a week of forensic analysis. The 14-Point gives everyone a common language: "you're failing #5 and #7" is a conversation; "your schedule feels off" is an argument. That objectivity is why it's now written into so many owner and federal schedule specifications — and why schedulers who consistently pass it get less scrutiny on everything else.

Make it a pre-submit habit

Run the 14-Point on every schedule before it leaves your desk — bids, baselines, monthly updates, claim narratives. Short on time? Start with the zero-tolerance three (#2 Leads, #7 Negative Float, #9 Invalid Dates) and the integrity test (#12); clean those and you've caught the problems that embarrass schedulers in front of owners. For the two findings that trip people up most, we wrote the field playbooks: Leads (#2) and Negative Float (#7).

The 14-Point won't write a good schedule for you. But in about two minutes, it will tell you whether anyone should believe the one you've got.