Philippines staffing research · Published: September 18, 2026

Can a dependency audit make project dates more credible?

A schedule-review protocol for missing links, constraints, progress evidence, and manager-owned forecasting decisions.

Key Stats

The U.S. Government Accountability Office Schedule Assessment Guide presents ten practices for reliable schedules. Its sequencing guidance explains that missing or incorrect dependencies can produce unreliable float and critical-path results.

Methodology

This desk review checked the GAO Schedule Assessment Guide and NIST information-quality guidance on September 18, 2026. It adapts those sources into an eight-week shadow audit for one approved project schedule. No client plan, budget, employee record, project system, or delivery result was accessed. The design evaluates schedule logic and evidence at a cutoff; it does not forecast a particular project or certify a completion date.

Key Takeaways

Research question. A project schedule can contain hundreds of dated tasks and still give a weak account of what drives completion. The study asks whether a Philippines-based project coordination lane can surface missing dependency logic without taking authority over scope or commitments. The unit is a detailed activity or milestone at a recorded schedule version. The project manager defines the approved schedule, calendars, status date, activity classes, and materiality rules. Summary bars, recurring administrative tasks, external milestones, and level-of-effort work need separate treatment rather than one generic test.

Guidance basis. GAO describes a reliable schedule as comprehensive, well constructed, credible, and controlled. Its sequencing practice calls for logical relationships between activities and warns that missing or incorrect dependencies can prevent early delays from propagating to dependent work. It also asks whether date constraints, leads, and lags are justified. These are source facts from a government assessment guide. The operational inference is that a coordinator can test observable schedule fields and prepare questions. Only the responsible manager can decide whether the modeled relationship reflects the real work.

Baseline and population. Freeze the approved schedule at the study start and again at each reporting cutoff. Build the population from detailed activities active or planned inside the review horizon. Preserve deleted and added activities through version comparison. Stratify by workstream, responsible owner, milestone proximity, constraint type, and external dependency. Review all activities on or near the calculated critical path plus a reproducible sample outside it. The protocol should state the scheduling software, calculation settings, calendars, and data date because those choices can change the apparent result.

Logic checks. Identify activities without appropriate predecessors or successors, links to summary activities, unusual relationship types, negative or unexplained lag, hard constraints, overdue starts, progress beyond the status date, and completion without accepted evidence. A finding is a question, not an automatic defect. Procurement, regulatory gates, external events, and fixed launch windows may justify constraints. The record should connect each exception to the owner’s explanation and final disposition. Recalculate only in a controlled copy until the project manager approves a change.

Progress evidence. Status should separate work started, work completed, deliverable accepted, and milestone approved. A percent-complete field without a defined measurement basis may hide uncertainty. Coordinators can gather the latest authorized source, record its observation time, and flag conflicts. They should not infer progress from calendar time or private messages. When a work owner cannot verify status, preserve “unverified” rather than carrying forward an old value as current. This makes uncertainty visible in forecast review instead of silently converting it to apparent precision.

Measures. Report missing-link candidates, justified exceptions, unjustified constraints, unverified progress, accepted corrections, and changes to forecast dates after manager approval. Show counts by activity class and the population denominator. Compare the distribution of total float and forecast movement before and after accepted corrections, while explaining that recalculation is mechanical, not causal proof. A schedule with fewer flags is not necessarily healthier; it may be smaller, less detailed, more constrained, or reviewed under weaker rules.

Authority boundary. Philippines-based project coordinators may maintain versions, run approved logic checks, request status evidence, prepare exception notes, and calculate scenarios in a copy. Project managers and accountable work owners retain scope, duration, sequencing, resource, baseline, risk-acceptance, and commitment decisions. A coordinator must not move a contractual milestone, mark a deliverable accepted, or change a dependency to improve the reported critical path. If owners disagree, preserve the disagreement and route it through the project’s governance path.

Analysis and decision use. Examine which exception types repeat by workstream and whether owner explanations arrive before the reporting cutoff. Use the findings to improve status requests, activity definitions, or review cadence. Do not use the audit as an individual performance score. A workstream with many findings may be better documented or undergoing more change. The study can support a decision about where to inspect and which modeling rule needs clarification. It cannot show that a team will meet the resulting date or that a chosen plan is optimal.

Worked interpretation. Consider a testing activity with a fixed start date but no predecessor, even though the project narrative says it depends on a completed build. The observed facts are the constraint, absent link, narrative source, schedule version, and status date. The analysis is that the model may not transmit a build delay to testing. A coordinator can prepare two controlled scenarios: the current model and a version with the proposed dependency. The project manager checks whether the narrative describes a real requirement, whether another activity already represents the gate, and whether the constraint has a documented reason. Only after approval does the live schedule change. The resulting forecast movement is a mechanical consequence of the accepted model, not proof that the later date will occur. If the manager rejects the link, retain the explanation so the same candidate does not reappear without context. This evidence makes the next audit more efficient and lets reviewers distinguish an intentional exception from an overlooked relationship. It also keeps schedule ownership with the manager rather than the person running the diagnostic.

Limitations and conclusion. Incomplete work breakdowns, hidden external schedules, resource constraints, uncertain estimates, software settings, and optimistic status reports can all limit the result. Logic quality is only one part of schedule credibility, and schedule credibility is only one part of project control. This protocol does not perform quantitative schedule-risk analysis. It provides a practical evidence trail from activity, relationship, and status source to an owner-approved correction. That narrow trail makes distributed project coordination more reviewable without giving administrative support control over promises.

FAQs

Sources

1. https://www.gao.gov/products/gao-16-89g
2. https://www.nist.gov/director/nist-information-quality-standards

Related Research

• philippines-project-decision-readiness-research
• philippines-project-status-reporting-research