The DCMA relationship types metric measures the proportion of finish-to-start (FS), start-to-start (SS), finish-to-finish (FF) and start-to-finish (SF) relationships in a schedule. Under the commonly used 14-point assessment criteria, FS relationships should account for at least 90 percent of the relationships associated with incomplete work.
However, 90 percent is a diagnostic threshold, not a command to convert every link to FS. A result below the threshold tells the scheduler to examine the network for excessive concurrency, weak activity decomposition, hidden lags and incomplete logic. Valid SS and FF relationships should remain when they accurately represent how the work will execute.
What the DCMA relationship types metric measures
The relationship types metric is one element of the broader DCMA 14-point schedule assessment. The Defense Acquisition University describes these metrics as a framework for identifying potential problem areas that merit further analysis. Therefore, a failed metric does not prove that the schedule is invalid.
The basic calculation is:
FS percentage = FS relationships ÷ total FS, SS, FF and SF relationships × 100
For example, assume an Integrated Master Schedule (IMS) contains 1,000 relationships associated with incomplete detailed activities. If 875 are FS, the schedule has an FS percentage of 87.5 percent. That result falls below the 90-percent assessment threshold.
The scheduler should also confirm how the analysis tool defines the population. Different tools or organizational procedures may treat completed tasks, milestones, external links and level-of-effort activities differently. Document the counting rules before comparing results across reporting periods or schedules.
The DAU Project and Program Management Fundamentals Handbook explains that FS relationships should represent at least 90 percent of the relationship types in the assessed population. It also describes the 14-point metrics as indicators of potential schedule problems rather than stand-alone conclusions.
The four schedule relationship types
A relationship defines which event on the predecessor controls an event on the successor. Microsoft Project supports four dependency types and uses FS as its default. The Microsoft guidance for linking tasks provides the software-specific definitions.
Finish-to-start
With an FS relationship, the predecessor must finish before the successor can start. For example, “Complete circuit card design” must finish before “Release circuit card drawings” can start.
FS logic usually provides the clearest representation of sequential work. It also makes paths easier to trace because the predecessor’s finish directly controls the successor’s start. As a result, duration changes move downstream work in a predictable way.
Start-to-start
With an SS relationship, the predecessor must start before the successor can start. The relationship does not require both activities to start on the same day. It only establishes the predecessor’s start as the earliest point from which the successor may start.
SS logic can model legitimate concurrent work. However, it may fail to transmit a late predecessor finish after both activities have started. In addition, an SS link alone does not establish what must occur before the successor can finish.
Finish-to-finish
With an FF relationship, the predecessor must finish before the successor can finish. The two activities do not have to finish simultaneously. Instead, the predecessor’s finish establishes the earliest allowable finish for the successor.
FF logic can represent a valid completion dependency. However, an FF predecessor alone does not control the successor’s start. The successor may effectively back into an earlier start based only on its duration and finish logic.
Start-to-finish
With an SF relationship, the predecessor must start before the successor can finish. This relationship is uncommon and often difficult for reviewers to interpret. The NASA Schedule Management Handbook identifies SF as rarely used and recommends using FS as often as possible.
An SF link may occasionally represent a valid handover, such as an outgoing support shift ending only after the replacement shift starts. Even then, the scheduler should confirm that the relationship direction and activity names communicate the intended sequence.
Why finish-to-start logic is preferred
FS logic tends to produce a more transparent critical path because it connects complete units of work in a clear sequence. The GAO Schedule Assessment Guide states that most relationships in a detailed schedule should be FS. GAO also warns that an overabundance of SS and FF relationships may indicate optimistic overlap or shortcuts in activity decomposition.
FS relationships offer several practical advantages:
- They make predecessor and successor paths easier to trace.
- They transmit duration changes through the network clearly.
- They reduce the risk of open starts and finishes within otherwise linked activities.
- They support more understandable critical and near-critical path analysis.
- They make logic reviews easier for control account managers, program managers and customers.
Still, the schedule should model the work rather than pursue a score. Replacing valid concurrency with artificial FS sequencing can extend the forecast and misrepresent the execution plan. Conversely, converting a questionable SS relationship to FS without revisiting the activity definitions may only hide the original modeling problem.
SS and FF logic can create dangling endpoints
The relationship type count does not reveal whether each activity has both its start and finish connected to the network. An activity may have predecessors and successors yet still contain dangling logic.
For example, an activity with only an FF predecessor has a controlled finish but no logical driver for its start. Likewise, an activity with only an SS successor may control another activity’s start without passing the effect of its own finish downstream.
GAO recommends checking whether each activity start, except the project start, has an FS or SS predecessor. Each activity finish, except an appropriate terminal milestone or deliverable, should also drive an FS or FF successor. This endpoint review goes beyond simply checking for tasks with blank predecessor or successor fields.
For a focused discussion of open-ended activities, see the DCMA logic check for missing predecessors and successors.
A realistic program example
Consider a fictional radar-upgrade program. Its IMS contains a 20-day “Conduct Environmental Qualification Test” activity and a 20-day “Monitor and Record Test Data” activity. The data-monitoring activity starts when the test starts and cannot finish before the test finishes.
The scheduler models two relationships:
- Environmental test SS to data monitoring.
- Environmental test FF to data monitoring.
These non-FS relationships represent the work accurately. The SS link controls the monitoring start, while the FF link connects its finish to completion of the test. Converting both links to a single FS relationship would incorrectly delay monitoring until testing had ended.
Elsewhere, the schedule contains “Develop Qualification Procedure,” linked SS plus 10 working days to “Review Qualification Procedure.” That structure assumes reviewers can begin after ten days, regardless of what content is available. It also hides the reviewable product inside a lag.
The team decomposes the work into “Draft Procedure Sections,” “Review Draft Sections,” “Incorporate Review Comments” and “Approve Qualification Procedure.” The revised sequence uses mostly FS logic and exposes measurable handoffs. It also removes the hidden delay and gives the control account manager better status visibility.
The result is not merely a better metric score. The network now reflects the products that move between engineering, test and quality organizations.
Relationship type and lag are separate issues
A relationship may include an offset. For example, FS plus five working days remains an FS relationship, but it also contains a lag. Therefore, it affects both the relationship types metric and the separate lag assessment.
Do not use lag to conceal identifiable work. If five days represent adhesive curing, government review or material conditioning, a discrete activity may provide better visibility. However, some lags represent the simple passage of time and may be justified. The deciding factor should be whether the model remains measurable, traceable and risk capable.
Review the rationale in when schedule lags are appropriate and the separate DCMA lags check. Negative lag, often called lead, creates additional concerns because it allows a successor event before the specified predecessor event. See why leads can damage schedule quality for that issue.
How to review a failed relationship types result
- Confirm the calculation. Verify the task and relationship population, exclusions and treatment of external dependencies.
- Group the exceptions. Separate SS, FF and SF links by work breakdown structure, control account, subcontractor or responsible organization.
- Review SF relationships first. They are rare and may indicate reversed predecessor-successor logic.
- Check SS and FF endpoints. Confirm that starts and finishes remain connected to valid logic.
- Inspect associated lags. Determine whether the offset represents hidden work or an unavoidable passage of time.
- Test the execution sequence. Ask the responsible technical owner what product or event permits the successor to start or finish.
- Decompose broad activities. Replace artificial overlap with discrete handoffs when the work supports that level of detail.
- Retain justified exceptions. Document valid concurrent logic rather than changing it solely to reach 90 percent.
- Recalculate the schedule. Review changes to the critical path, total float, milestone forecasts and resource demand.
Finally, trace key completion paths after making logic changes. A higher FS percentage does not guarantee a valid network. The schedule must still respond correctly when driving work slips. The article on tracing a critical path in Microsoft Project explains how to verify the resulting path.
Is the 90-percent FS threshold contractual?
The 90-percent threshold is not automatically a contractual requirement. The text of DFARS 252.234-7002 requires an applicable contractor to use a compliant Earned Value Management System and generate timely, reliable and verifiable information for required reports. It does not prescribe a 90-percent FS ratio.
A solicitation, contract, Contract Data Requirements List, data item description, program directive or approved scheduling procedure may establish additional expectations. Those requirements can also vary by agency, program and contract tailoring. Therefore, teams should distinguish three separate questions:
- Does the schedule meet the contract’s explicit requirements?
- Does it comply with the organization’s approved scheduling procedures?
- Does the relationship profile indicate a technical weakness that management should investigate?
The DCMA relationship types metric mainly addresses the third question. Use it as a screening tool, then evaluate the underlying network before directing corrective action.
Practical conclusion
A strong detailed schedule will usually contain mostly FS relationships because discrete work often moves through identifiable handoffs. However, the correct objective is not an FS-only network. The objective is complete, dynamic logic that represents how the program will execute.
Use the 90-percent threshold to locate possible weaknesses. Then review non-FS relationships for valid concurrency, endpoint logic, hidden work and unnecessary offsets. When an SS or FF relationship accurately models execution, retain and document it. When it hides an undefined handoff, improve the activity structure instead of merely changing the link code.