What Is the DCMA-14? A Complete Guide to Schedule Quality Assessment

The DCMA-14 schedule assessment is a set of 14 schedule-quality checks (missing logic, leads, lags, relationship types, hard constraints, float and duration limits, invalid dates, resources, missed tasks, a critical path test, CPLI and BEI) published in the Defense Contract Management Agency's October 2012 EVMS Program Analysis Pamphlet. It exists to find problem areas in a contractor's integrated master schedule. Most thresholds are 5%, and DCMA treats them as flags, not automatic failures.

Where do the 14 checks come from?

The checks are Section 4, "14 Point Schedule Metrics for IMS," of DCMA-EA PAM 200.1, the EVMS Program Analysis Pamphlet dated October 2012. The text says they were "developed to identify potential problem areas with a contractor's IMS." Mosaic Projects describes that pamphlet as the last published version. Vendors such as Deltek date the assessment to 2005 and say it has spread beyond defense work into commercial projects. That origin date comes from vendors, not from the pamphlet.

Deltek also calls the metrics "advisory guidelines rather than a compliance mandate." That wording is worth remembering whenever an owner's specification says a schedule "must pass DCMA-14."

Note one more detail about scope. The pamphlet's analysis population "should exclude Completed tasks, LOE tasks, Subprojects (Summary tasks in MS Project), and Milestones." So when a check says "% of incomplete tasks," it means incomplete detail tasks only. Software tools count this differently, and Mosaic's white paper notes the implementations are not certified by DCMA or any other body.

What does the DCMA-14 schedule assessment check, and what are the thresholds?

The table below uses the values in the pamphlet. Where a secondary source differs, the individual metric articles say so.

Two points trip people up. Leads and lags are measured against logic links, not tasks, and Relationship Types is also a link count. The percentage metrics with a 5% limit are measured against incomplete tasks, except Missed Tasks, which uses tasks due by the status date.

How the 5% limits are counted

The denominator matters as much as the threshold. Suppose a schedule has 400 activities, of which 120 are complete, 30 are level-of-effort or summary rows, and 40 are milestones. Under the pamphlet's population rules, the base for the Logic, Hard Constraint and High Float checks is the remaining 210 incomplete detail tasks, so 5% is about 10 tasks. A tool that counts all 400 activities would allow 20 and report a better score for the same file.

Links work the same way. If the network has 600 logic links, 5% for lags is 30 links, and Relationship Types needs at least 540 Finish-to-Start links. Because mixed-in milestones and summaries shift every ratio, always state the population when you report a result.

How do the first six structure checks work?

Metrics 1 through 6 test whether the network is built so the dates mean something. They share one pattern: each one finds a way a schedule can look fine while the logic underneath no longer drives the dates.

Logic. Every incomplete task should have at least one predecessor and one successor. See the logic metric guide for the P6 filter approach.

Leads and lags. Leads are negative lags, and DCMA's goal is zero. Lags are capped at 5% of links. Both distort float. Why leads are flagged and when a lag is valid cover the construction cases.

Relationship types. At least 90% of links should be Finish-to-Start. The relationship types article explains why start-to-start links draw attention and how to justify them.

Hard constraints. DCMA lists Must-Finish-On, Must-Start-On, Start-No-Later-Than and Finish-No-Later-Than as hard because they "may prevent tasks from being moved by their dependencies." P6 draws the line differently, which the hard constraints article untangles.

High float. Total float over 44 working days (two months) on more than 5% of incomplete tasks suggests the network "may be unstable and may not be logic-driven." See high float for causes and the debate over a fixed cutoff.

What do the remaining metrics test?

Metrics 7 through 14 move from structure to status and performance.

Negative float flags any incomplete task with total float under zero. The pamphlet gives no percentage but says each one "should have an explanation and a corrective action plan" and that ideally there should be none. The negative float guide covers P6 causes.

High duration applies only to baseline durations over 44 working days for tasks starting in the detail planning period. Long placeholder activities in later rolling-wave periods are not counted. See high duration activities.

Invalid dates catches schedules that were not properly statused. In P6, this usually traces to the data date and out-of-sequence settings. The invalid dates article walks through it.

Resources is the most misread check. DCMA's IMS data item description (DI-MGMT-81650) "does not require the contractor to load resources directly into the schedule," so the check applies only when a contractor does. Resource requirements are usually set by the contract, not DCMA.

Critical path test, CPLI and BEI are the performance checks. Introduce an intentional slip on the critical path and the completion date should move in direct proportion; if not, "there is broken logic." CPLI and BEI are explained with worked arithmetic in the BEI and CPLI guide.

Worked example: the critical path test

The critical path test is the one check that cannot be read from a report. You change the schedule and watch what happens. Illustrative example, with assumed durations: a four-activity driving path ends at Substantial Completion.

The end milestone should show about 20 working days of negative float, or an early finish 20 days later. If Substantial Completion does not move, a successor link is missing somewhere downstream of A1030, which is the "broken logic" the pamphlet describes. Run the test on a copy of the schedule, never the working file.

What are the formulas for CPLI and BEI?

Both indexes are ratios where 1.00 means on pace.

• CPLI = (critical path length + total float) ÷ critical path length, with length measured in work days from the status date to the milestone.
• BEI = cumulative tasks completed ÷ cumulative tasks with a baseline finish on or before the status date.

Illustrative example: assume a critical path of 100 working days to a milestone with total float of -10. CPLI = (100 + -10) ÷ 100 = 0.90, which is below 0.95 and flagged. Now assume 100 tasks were baselined to finish by the status date and 90 are complete. BEI = 90 ÷ 100 = 0.90, also flagged. These numbers are made up to show the arithmetic.

BEI ignores whether a task finished early or late. The pamphlet pairs it with a Hit Task percentage: tasks completed on or before their baseline date divided by tasks with a baseline finish in the current fiscal month.

A common error is quoting 1.00 as the BEI or CPLI threshold. Several vendor guides, including SmartPM's summary, present 1.0 as the target. The pamphlet's flag is 0.95.

Is the DCMA-14 still used, and is it pass or fail?

On the second question, DCMA's own text is clear. The metrics "provide the analyst with a framework for asking educated questions and performing follow-up research." A red metric "is not in and of itself synonymous with failure," and "correction of that metric is not necessarily required, but it should be understood."

On the first, the picture is mixed. A September 2022 NDIA working-group slide, co-presented by a Raytheon representative and a DCMA representative, says "14-pt is still evident in contract CDRLs," "No longer used by DCMA (EVMS Center)," and that certain analytics are still used to confirm schedule quality. No newer DCMA publication was found. Vendors continue to present the assessment as current, so verify what a given customer actually requires.

What does a schedule have to do to get approved?

The assessment itself does not approve anything. Approval comes from the contract or specification. Here is what to do in practice, as of 2026.

1. Read the contract clause. Does it name DCMA-EA PAM 200.1, a tool and version, or a subset of metrics? Many specifications say only "DCMA-14," which leaves the population and thresholds undefined.
2. Run the checks in the tool the reviewer will use, because counts vary with the denominator each tool uses.
3. For each red metric, write a short explanation of cause and a corrective action if one is warranted. For negative float, the corrective action plan is explicit in the pamphlet.
4. Fix the structural causes first. Missing logic often produces high float and a failed critical path test at the same time, so one repair clears several metrics.
5. Keep the explanation with the submittal. A reviewer who sees a flagged lag with a documented curing period will usually move on.

GAO takes a broader view. Its Schedule Assessment Guide, GAO-16-89G, issued December 2015, lists ten best practices, from capturing all activities and sequencing them to a valid critical path, reasonable total float, a schedule risk analysis and maintaining a baseline. Those cover ground the 14 checks do not, such as risk and traceability.

Who uses the DCMA-14, and how do they score it?

Three groups use the checks, and each reads them differently.

Defense and federal contractors run the assessment on the integrated master schedule because a contract data requirement often calls for it. The pamphlet was written for analysts reviewing earned value systems, so its language assumes a single integrated schedule tied to a baseline.

Owner-side reviewers on large commercial and public projects borrow the checks as a screening list. They typically run the contractor's native file through a scheduling analytics tool or the built-in checks in P6, compare the results with the thresholds above, and ask for explanations of anything flagged. Because tools count differently, the reviewer's tool and version are part of the answer. Owners who want consistent results should name them in the specification.

Contractors and consultants use the same checks before submitting, so that nothing in the file surprises the reviewer. A pre-submittal run also gives you the time to repair root causes instead of writing excuses.

Scoring is not standardized. Some tools report each metric as pass or fail against the pamphlet's limit, and some roll the 14 into an overall percentage, often weighting metrics differently. Those roll-ups are vendor features, not part of the DCMA pamphlet, so a score of 92% from one tool does not translate to another. Report the individual metrics and the populations behind them, and treat any composite score as a convenience.

What are the common mistakes when applying the DCMA-14?

Most failed reviews trace to a handful of avoidable errors:

• Treating the thresholds as a contract requirement. The pamphlet does not bind a construction contractor. The specification does, and a clause naming "DCMA-14" without a tool, version or population is open to dispute.
• Using 1.00 as the BEI or CPLI limit. A schedule with a BEI of 0.97 is not flagged by the pamphlet.
• Reading Resources as mandatory. The check only applies to loaded schedules.
• Counting milestones and summaries. They sit outside the pamphlet's population and will inflate or deflate your percentages.
• Fixing the number, not the cause. Deleting a constraint but keeping the forced date, or swapping a lag for a dummy activity, passes the check and keeps the problem. Reviewers who trend metrics across updates spot this quickly.
• Skipping the narrative. Because DCMA wants flags understood, a one-paragraph explanation often matters more than a green cell.

A passing score also says little about whether durations are realistic, the scope is complete, or calendars fit the work. The checks test structure and status, which is why GAO-16-89G adds practices such as schedule risk analysis.

Guides in this series

Each article below takes one part of the assessment in depth.

• Logic metric: missing predecessors and successors.
• Leads and lags: link-level timing offsets.
• Relationship types: FS, SS, FF and SF.
• Hard constraints: DCMA's list versus P6's.
• High float and negative float.
• High duration, invalid dates and resources.
• BEI and CPLI performance indexes.
• What a DCMA-compliant schedule really proves and how owners use the checks.