Guide · Schedule quality

DCMA 14-point schedule check explained

The DCMA 14-point assessment is the most widely used quick test of whether a CPM schedule is built well enough to be trusted. It does not tell you whether the plan is achievable — it tells you whether the network behaves like a network. This guide walks through each check, the usual threshold, what a failure actually means, and how to move beyond the stock 44-day limits to thresholds and weightings that suit your project type.

Origin

Developed by the US Defense Contract Management Agency to screen contractor schedules before deeper review.

What it tests

Network mechanics: logic, leads and lags, constraints, float, durations and progress realism.

One of several standards

TideLine runs the same engine against DCMA, GAO, AACE RP-90R-19, NDIA PASEG and USACE profiles — plus your own.

Tunable

Weight the metrics that matter, edit warn/fail limits, and set float and duration day limits to the project type.

The 14 checks and their usual thresholds

1. Logic — every activity except the first and last should have a predecessor and a successor. Threshold: no more than 5% missing logic. Open ends let activities float free of the network, so the critical path stops being meaningful.

2. Leads (negative lag) — target 0%. A lead hides real sequencing and distorts the forward pass. Replace it with a properly broken-down activity.

3. Lags — no more than 5% of relationships. Lags are often used to fake a duration that should be a visible activity, such as cure time or approval time.

4. Relationship types — at least 90% Finish-to-Start. Heavy use of SS/FF pairs makes the network hard to audit and can create hidden loops of dependency.

5. Hard constraints — no more than 5%. 'Must finish on' and 'mandatory' constraints override logic, which is exactly what you do not want in a live programme.

6. High float — no more than 5% of incomplete activities with total float above 44 working days. Very high float usually means missing successor logic rather than genuine slack.

7. Negative float — target 0%. Negative float means the network cannot meet a date you have imposed on it; it is a signal to re-plan or re-baseline, not something to leave sitting in an update.

8. High duration — no more than 5% of incomplete activities longer than 44 working days. Long activities hide progress and make earned value meaningless at activity level.

9. Invalid dates — no actual dates in the future and no forecast dates in the past relative to the data date. This is the single most common sign an update was not statused properly.

10. Resources — every activity with a duration should carry resources or costs where the schedule is resource-loaded.

11. Missed tasks — no more than 5% of activities finishing later than their baseline finish. This measures execution, not build quality.

12. Critical path test — deliberately add a large delay to a driving activity and confirm the project finish slips by the same amount. If it does not, the network is broken somewhere.

13. Critical path length index (CPLI) — 1.00 or above is on track; below 0.95 signals erosion of the path to the contractual completion date.

14. Baseline execution index (BEI) — 0.95 or above. Below that, the team is completing fewer activities than the baseline planned by this point.

How to read a failing check

A red metric is a prompt to look, not a verdict. Hard constraints at 8% on a schedule with contractual sectional completion dates may be entirely legitimate. Lags at 12% on a design programme with statutory approval periods may be defensible if each lag is documented.

The failures that should always be fixed are open logic, invalid dates, negative float that nobody has explained, and a failing critical path test. Those four mean the calculated dates cannot be relied on at all — everything downstream, including earned value and delay analysis, inherits the error.

DCMA is one standard among several

DCMA is the baseline, not the whole picture. TideLine's Quality module runs the same underlying checks against five threshold profiles, because different reviewers judge the same network differently: DCMA 14-Point (balanced, 14 standalone checks), GAO Schedule Assessment Guide (auditor-grade — 100% logic expected, constraints must be justified, plus traceability and description-quality checks), AACE RP-90R-19 (forensic readiness — out-of-sequence work and constraint classification weighted heaviest), NDIA PASEG (EVM-aligned — cost loading, LOE discipline, CPLI and BEI elevated) and USACE EM-1110-2 / NAVFAC (construction and federal infrastructure — long durations expected, resource loading required, hardest on constraints).

Because the engine is shared, you can assess the same update against two profiles and see exactly which findings are standard-specific rather than genuine defects. A schedule that passes DCMA on hard constraints at 4% can still fail GAO or USACE, which weight constraints far more heavily.

Weighting metrics and setting your own bad-quality limits

The overall score is a weighted result, not a simple count of red lights. Each metric carries a weight, and the weights differ per standard: DCMA gives logic, negative float, invalid dates and the critical path test a weight of 3.0 while treating resources as an optional 0.5 check; AACE pushes the critical path test and out-of-sequence work to 4.0; NDIA lifts CPLI and BEI to 2.5; USACE puts hard constraints at 4.0.

You can override all of it. In Quality settings, edit the warn and fail thresholds per metric, change the weight so the metrics that matter on your contract dominate the score, disable checks that do not apply, add your own custom criteria, and save the result as a named custom profile you reuse on every update.

The duration and float limits are parameters, not constants. DCMA's 44 working days for high float and high duration suits programme-level defence work; AACE uses 60 days for industrial schedules; USACE allows 60 days of float and 90-day durations because a concrete or earthworks activity legitimately runs that long. Set the day thresholds to your project type — fit-out and shutdown schedules usually want tighter limits than 44 days, while linear infrastructure needs looser ones — rather than reporting failures the work never deserved.

Running the check on a P6 schedule

Export the update as an .xer file and load it into TideLine's Quality module, or assess the schedule you built in TideLine directly. The metrics are calculated against the loaded network with the offending activities and relationships listed underneath each one, so you can jump straight to the 30 activities driving a failing lag percentage rather than hunting through a filter in P6.

Runs are saved with a snapshot of the profile and thresholds used, so a later review can see exactly what standard produced a score. The Compare tab puts two loaded datasets, or two saved runs, side by side — the practical way to show that quality is improving update on update rather than arguing about a single number.

Because the same project feeds the Metrics, Delay and Reports modules, a monthly quality run becomes part of the update cycle: check the network, then read float trend, earned value and critical path movement on a network you know is sound.

Frequently asked questions

Is the DCMA 14-point check mandatory?
Not generally. It originated as a US government screening tool, but many owners, PMCs and contractors now adopt it voluntarily as a schedule quality standard because the thresholds are published and easy to test.
Which schedule quality standards does TideLine support?
Five profiles run on the same engine: DCMA 14-Point, GAO Schedule Assessment Guide, AACE RP-90R-19, NDIA PASEG and USACE EM-1110-2 / NAVFAC. Each brings its own thresholds, weights and standard-specific checks, and you can save your own custom profile on top.
Can I weight one quality metric more than another?
Yes. Every metric has an editable weight that drives the overall score, alongside editable warn and fail thresholds. You can also disable checks that do not apply and add custom criteria, then save the configuration as a reusable profile.
Can I change the 44-day float and duration limits?
Yes — they are parameters per profile. AACE uses 60 days, USACE allows 60 days of float and 90-day durations for construction, and you can set your own values to match the project type so long, legitimate activities are not flagged as defects.
What is a good overall score?
There is no official composite score. Most reviewers expect the logic, invalid date, negative float and critical path tests to pass outright, and treat the percentage-based metrics as items requiring explanation rather than automatic failure.
What is the difference between CPLI and BEI?
CPLI measures whether the critical path still reaches the required completion date — it is forward-looking. BEI measures how many baseline activities have actually been completed by now — it is backward-looking execution performance.
Can I run DCMA checks without Primavera P6?
Yes. TideLine reads P6 .xer exports in the browser, rebuilds the CPM network and runs all 14 checks, so you do not need a P6 licence on the machine doing the review.