Esc

↑↓ move↵ openIndex · Pagefind

Time impact analysis: build a model another reviewer can follow

Prepare a traceable time impact analysis with a documented base schedule, event fragnet, controlled calculation and clear limits on the conclusion.

Charley Forey2 min readSchedulingSchedule forensics
On this page
  1. Keep the comparison controlled
  2. Make the fragnet inspectable
  3. Explain why event duration and finish movement differ
  4. Provide a review package

A time impact analysis examines an event through a schedule model. Its usefulness depends on the base schedule, the event logic and the assumptions linking the event to the remaining work. An unexplained difference between two finish dates is not enough for another person to review the result.

Start with the project’s requirements for notice, timing, submissions and analysis method. Do not assume every specification uses the term “time impact analysis” in exactly the same way. The SCL Delay and Disruption Protocol provides further discussion of prospective analysis and delay-management practice.

Keep the comparison controlled

From event to reviewable result
  1. Identify the event and the question.
  2. Validate and preserve the selected base update.
  3. Model the event and its connections.
  4. Calculate a separately identified scenario.
  5. Compare milestones and driving paths.
  6. Explain assumptions, evidence and limitations.

Confirm the base schedule’s data date, progress, calendars, logic and calculation settings. Record any corrections before adding the event. If the unimpacted copy differs from the received file, disclose why and retain both.

Make the fragnet inspectable

A fragnet is a small network representing the event and its connection to the project. Give its activities stable IDs, understandable descriptions, supported durations, calendars and predecessor/successor ties. Identify which existing work must wait and what physical or administrative condition releases it.

For a fictional design revision, the model might include revised information, review, fabrication adjustment and delivery. Those activities need evidence; the model should not simply insert an arbitrary delay into the completion milestone.

Explain why event duration and finish movement differ

Suppose an event adds five working days to a path with two days of available path margin. In a simple single-calendar network, with unchanged logic and no competing changes, the finish could move three working days. The five-day event and three-day movement describe different quantities.

02 / MODEL EFFECTEvent duration ≠ finish movement.

Read the network: Different calendars, competing paths, or revised logic can change the result.

Illustrative single-calendar model: a five-working-day event uses two working days of available path margin, leaving a three-working-day finish movement. This assumes unchanged logic and no competing changes; it is not a universal time-impact formula.

A real model may behave differently because another path becomes controlling, work overlaps, calendars differ or assumptions change. Trace the result through the actual network rather than applying the example as a universal subtraction rule.

Provide a review package

Include the base identifier, event evidence, fragnet activity list, calculation settings, before/after milestone table, path explanation and assumption register. Keep timing clear: a forecast made during the project is different from an analysis made later using historical knowledge.

Label the calculated movement as the model’s result. Responsibility, notice compliance, mitigation and entitlement require their own evaluation. Where input evidence is incomplete, say which part of the result depends on the assumption.

Next, compare delay methods, prepare extension-of-time evidence, or follow the full Schedule Forensics learning path.

Share this article

Written by

Charley Forey

Founder of os.construction. We're building the operating system for construction. Project Financials follows job money, and company Accounting covers the books and controls. AI agents remain planned. We write about how money and data actually move on a job, and we build in public.

Apply to the founding program →Start a free trial →
Open file