What Is Forensic Delay Analysis, and When Do You Actually Need One

Your project finished five months late. The owner is withholding liquidated damages. Your subcontractor just filed for an extension of time. Everyone at the table says it's someone else's fault, and nobody can prove it because nobody can explain exactly what happened to the critical path, when it happened, and who caused it.

You need a forensic delay analysis. Not a schedule review. Not a health check. Not a letter from your scheduler saying "it wasn't our fault." You need a structured, methodology-driven reconstruction of the project timeline that can survive cross-examination in an arbitration hearing.

This is the analysis that determines who owes whom: time, money, or both. Here's what it actually is, when you need one, and what to look for in both the deliverable and the analyst.

What Forensic Delay Analysis Is (and Isn't)

A forensic delay analysis is a retrospective, evidence-based examination of what happened to a project's critical path over time. It compares what was planned against what actually occurred, identifies the events that caused the schedule to slip, attributes those delays to the responsible parties, and quantifies the time impact.

It is not the same as a schedule health check. A DCMA 14-point assessment tells you whether your schedule is well-built. A forensic delay analysis tells you who broke it and by how much. The health check is the physical exam. The forensic analysis is the autopsy.

It is not an opinion letter. "In my professional judgment, the owner caused the delays" doesn't survive five minutes of cross-examination without structured methodology behind it. Arbitrators and judges want to see the math: every window, every critical path shift, every concurrent delay attribution. A narrative with a conclusion stapled to the end will not do.

The Nine Methods in AACE RP 29R-03

AACE International Recommended Practice 29R-03, Forensic Schedule Analysis, sorts the accepted methods into nine method implementation protocols, or MIPs. The main split is between observational methods, which study the schedules as they are, and modelled methods, which insert delays into a schedule or take them out and recalculate it.

The method is chosen for the records that exist and the question being asked, not by preference. CPP works to AACE RP 29R-03, RP 52R-06 and the SCL Delay & Disruption Protocol, 2nd Edition.

The Three Methodologies You Need to Know

Of the nine, three cover the vast majority of construction delay disputes. Each one answers a different question, and a thorough analysis often uses more than one.

1. Windows Analysis (AACE MIP 3.3)

This is the workhorse. The project timeline is divided into discrete periods, called "windows", typically aligned with schedule update cycles. Within each window, the analysis identifies what happened to the critical path: which activities slipped, which new delays appeared, and who was responsible for each one.

The power of windows analysis is its contemporaneous, period-by-period granularity. Instead of looking at the whole project and saying "it was late," you're looking at each month and saying "in this window, the critical path shifted from structural steel to electrical rough-in because the owner's RFI response was 22 days late, causing a 14-day net delay to the completion milestone."

Windows analysis is the primary method for most delay claims because it handles concurrent delay attribution: situations where both the owner and contractor caused delays in the same period. The conservation rule is an equality across the whole evaluation rather than a ceiling on any one window: AACE RP 29R-03 §3.3.E.13 sums the net gains and losses for each period to arrive at the overall impact, so the attributed totals reconcile to the observed project drift, recovery periods included.

2. Collapsed As-Built / But-For (AACE MIP 3.8)

This is the cross-check. The method takes the as-built schedule and systematically removes one party's delays, then recalculates the project completion date. The difference between the as-built finish and the "but-for" finish is the delay attributable to the removed party.

If opposing counsel is going to challenge your windows analysis, they're going to demand a but-for as a cross-check by a second method. Having both methods in your submission, and showing that they produce consistent results, significantly strengthens the claim.

3. Time Impact Analysis (AACE MIP 3.7 / RP 52R-06)

TIA models a specific delay event. It inserts a "fragnet" (a fragment of network logic representing the delay) into the CPM schedule and recalculates to see how it moves the completion date.

Done while the work is under way, under AACE RP 52R-06, it quantifies the impact of a specific change order, RFI delay or owner-directed change before the project finishes, typically as part of an extension of time request. MIP 3.7 in RP 29R-03 applies the same additive modelling to a series of schedule updates, one after another.

When You Need a Forensic Delay Analysis

Not every late project needs a forensic analysis. Sometimes the reason is obvious and everyone agrees. But if any of the following are true, you need one:

Liquidated damages are being assessed or threatened. If the owner is withholding money for late completion, the only way to demonstrate that the delays were not your responsibility is a methodology-compliant forensic analysis. "We were delayed by the owner" is a position statement. A forensic analysis makes it evidence.

You're submitting an Extension of Time (EOT) claim. Most contracts require the contractor to demonstrate the impact of owner-caused delays on the critical path. A schedule narrative without forensic methodology behind it will be rejected by any competent owner's consultant.

You're heading into arbitration or litigation. If the dispute has escalated to a legal proceeding, you need an analysis that can be presented as expert evidence. That means recognized methodology, traceable data, and an analyst who can be examined and cross-examined on their findings.

A subcontractor is filing a backcharge or delay claim against you. You need your own forensic analysis to evaluate their claim. Taking their word for the delay duration and impact, or counter-arguing without your own analysis, is bringing a narrative to a math fight.

The owner is withholding release of holdback. Some contracts tie holdback release to substantial completion. If there's a dispute about whether delays were excusable, the forensic analysis is what unlocks the holdback.

The Records a Forensic Delay Analysis Needs

The analysis is only as good as the record behind it. Before any method is chosen, the analyst needs three things:

• The schedules as native files. The baseline and every update as Primavera P6 XER files, not PDF prints. A PDF shows the dates; the native file holds the logic, the calendars and the constraints that produced them.
• The contract's time provisions. The notice, extension of time and delay clauses, which decide what the analysis has to show.
• The project record. Daily reports, RFIs and their responses, change orders and directives, notices of delay, meeting minutes and correspondence. The schedule shows when the work moved; these records show why.

If the schedule files are all you have, start there. Through the free claim check, one analyst reads them and says what they can support within a few business days, at no fee and with no obligation to retain, before anyone commits to a full analysis.

What a Good Deliverable Looks Like

A forensic delay report has usually been a static document: a narrative, some Gantt chart screenshots, and a summary table at the end. The reader can study it, but cannot test it.

Here's what the deliverable should actually include:

Interactive forensic dashboard. A browser-based HTML report where you can drill into each analysis window, see which activities shifted, trace the critical path movement, and explore the delay attribution at the activity level. This is what you screen-share in a mediation session when opposing counsel challenges a specific window.

Per-window and per-activity CSV exports. The raw data behind every conclusion, in a format that can be independently verified. If your analyst can't hand over the underlying data in a structured format, ask why.

Executive summary report. A DOCX narrative that walks through the findings in plain language, suitable for a project executive, a construction lawyer doing intake, or an arbitrator reading the submission for the first time.

Claims-ready packaging. Cover letter, per-event delay exhibits, supporting documents index, and mitigation documentation. A forensic analysis that produces findings but doesn't package them for submission is only doing half the job.

Audit trail. Every assumption, every data source, every methodology decision is documented and traceable. The opposing analyst is going to ask "how did you calculate the 14-day delay in Window 7?" and the answer needs to be in the deliverable, not in your analyst's head.

What to Look for in a Forensic Analyst

Not every scheduler is a forensic analyst. Scheduling is a technical skill. Forensic delay analysis is a technical skill plus the ability to produce evidence that survives a legal proceeding. Here's what separates the two:

Opinions built to be tested. Can your analyst defend every number in the report under hostile questioning? Ask how each figure traces back to the schedule data and the method named in the report.

Methodology compliance. The analysis should explicitly reference and follow a recognized AACE methodology. "I used my professional judgment" is not a methodology. MIP 3.3, MIP 3.7, MIP 3.8, RP 29R-03 and RP 52R-06 are methodologies.

Audit trail discipline. Every calculation should be reproducible by an independent analyst working from the same data. If your analyst can't explain exactly how they arrived at "22 days of owner-caused delay in Window 4," the opposing analyst will.

Conservation rule compliance. Across the evaluation, the attributed totals must reconcile exactly to the observed project drift: an equality per AACE RP 29R-03 §3.3.E.13, which sums the net gains and losses of every period, recovery windows included, to arrive at the overall impact. An analysis that assigns 15 days to the owner and 12 days to the contractor in a window where the project only slipped 18 days has a math problem. Opposing counsel will find it.

Cross-examination readiness. The deliverable should be written with the assumption that every sentence will be read aloud in front of an arbitrator and challenged. If the report contains hedging language, unsupported conclusions, or attribution without traceable evidence, it will not survive the hearing.

Counsel retaining a forensic scheduling expert will find the engagement model and the deliverables on the page for lawyers.

Forensic Delay Analysis in Ontario

Three things shape the work on an Ontario project:

• The contract sorts delays by cause. On a CCDC 2 contract, GC 6.5 sorts each cause of delay into a paragraph with its own consequences, so the analysis has to work event by event. See GC 6.5 of CCDC 2 - 2020, and How a Schedule Analysis Maps onto It.
• Adjudication runs on a short clock. Under the Construction Act, adjudication is a remedy used while the work is under way, and the party giving the notice has five days after the adjudicator agrees to conduct the adjudication or is appointed, to deliver the documents it intends to rely on. For that party, the analysis has to be finished before the notice goes out. See Prompt Payment and Adjudication Timelines, and What They Do to a Delay Claim.
• The expert has to be independent. Canadian courts screen expert evidence for independence and impartiality under White Burgess, and the consultant who built the schedules has a stake in them. See White Burgess, and How to Instruct a Schedule Expert.

The Cost of Waiting

Forensic delay analysis is most effective when engaged early, ideally as soon as delay events begin accumulating, not after the project is finished and the claim is overdue. Early engagement means the analyst can advise on contemporaneous record-keeping, flag critical path shifts in real time, and build the evidentiary record as the project progresses. If the same analyst is later asked to give expert evidence, the report says so, and the client is told beforehand that a court or tribunal may give that evidence less weight than evidence from an expert who had no part in preparing the claim.

Waiting until the project is over and the lawyers are involved costs more. The data is stale, the records are incomplete, and the analyst has to reconstruct months or years of project history from whatever documentation survived. That reconstruction takes longer, costs more, and produces a weaker result than an analysis built on clean contemporaneous data.