M6.5 — Claims & Forensic Analysis

M6.5 — Claims & Forensic Analysis

1. Practical outcome

The learner should be able to preserve relevant records, define a forensic question, establish a reliable chronology, test cause and effect across schedule and cost, distinguish facts from assumptions and produce reproducible findings. They should understand how project-controls evidence supports claims assessment, dispute prevention and learning, while recognising where contractual interpretation, legal advice or specialist delay/cost expertise is required.

M6.5 uses the baselines, actual records, forecast snapshots, change decisions and assurance evidence developed throughout Modules 1–6. It does not teach that a delay automatically creates compensation, that an incurred cost proves entitlement, or that one software method resolves every dispute. The practitioner provides a defensible account of what happened, its effects and the limits of the evidence, so authorised specialists and decision-makers can address the relevant commercial questions.

Ask: “What can the records demonstrate about this event and its consequence, what depends on an assumption, and could another analyst reproduce the result?”

2. Core concepts

Separate factual analysis, entitlement and valuation

A claim is a request for a contractual or other remedy. Factual analysis establishes events, instructions, performance and effects. Entitlement addresses whether the applicable contract and legal framework provide the remedy, including relevant procedures, notice requirements, responsibility and any exclusions. Valuation addresses the amount or time adjustment supported under that framework.

These questions interact but need separate reasoning. A late access event may be documented without establishing a compensable delay. A recorded cost may be genuine but unrelated, already allowed elsewhere or recoverable only under particular terms. A controls report should not silently turn a technical effect into a conclusion that a party owes a stated sum.

Preserve evidence before selecting a method

Relevant records include approved baseline and update files, calendars, logic and settings, actual dates, progress/acceptance evidence, access logs, instructions, notices, meeting records, resource diaries, time sheets, invoices, accruals, test results and correspondence. Preserve originals, metadata and the versions used at the time; distinguish analyst copies and transformations.

Use controlled access, retention and handling arrangements appropriate to the records and dispute context. Obtain specialist direction where preservation obligations, confidentiality or privileged material may be involved. Do not delete inconvenient versions, overwrite contemporaneous updates or circulate legal advice as ordinary project evidence. A record of custody and transformation helps explain how an original became an analytical input.

Define the question and time perspective

State the event, affected obligation, claimed remedy, analysis period and reference. A prospective assessment asks what was expected from the information available at the time. A retrospective assessment asks what actually happened using later evidence. Both can be useful, but mixing them can produce hindsight presented as a contemporary forecast.

Define the data date and the information available at each window. Later knowledge may help establish as-built facts; it must not be inserted invisibly into a model presented as the original prediction. Compare contemporaneous assessments with actual outcomes and explain differences rather than deleting the earlier forecast.

Chronology is necessary but not sufficient

An event register should connect dates, actors, instructions, activities, cost records and source references. Contradictory dates require investigation, not a choice of whichever supports the preferred position. Distinguish an event's occurrence from its discovery, instruction, effect and later resolution.

Temporal sequence alone does not establish causation. “Access was late and completion was late” needs an explanation of the work prevented, its controlling role, available alternatives, actual resources and downstream consequences. Identify other events that could have affected the same result and test them.

Delay to an activity, milestone and project are different

A late activity can use float without moving completion. A milestone can move while another path remains controlling. Paths can change over time as progress, access, resources, tests and interfaces develop. Explain the affected network and calendar rather than equating the event's duration with project delay.

Select a schedule-analysis approach that fits the question, contractual requirements, available records, model quality and timing. A contemporaneous time-impact assessment can insert a defined event into a credible statused model. A windowed retrospective review can examine actual progress and changing paths over successive periods. As-built comparisons and counterfactual models may help in suitable circumstances but require careful reconstruction and transparent assumptions. No label guarantees a reliable result.

Model quality and counterfactual assumptions

Check the baseline and updates for credible logic, calendars, actuals, constraints, open ends, out-of-sequence treatment and significant revisions. Correcting an unreliable model may be necessary, but preserve the original and disclose each repair and its effect. Do not rebuild the network solely to obtain a desired delay result.

A counterfactual asks what would have happened under a defined alternative, such as removing one event while retaining other observed conditions. It is a hypothesis, not an observed fact. State assumptions about resources, access, productivity, approvals and mitigation, and test sensitivity. A result that depends strongly on an unsupported assumption needs qualification.

Overlapping events and concurrency

Two events happening during the same dates do not necessarily create concurrent delay to completion. Examine whether each was an effective cause of the relevant completion effect under the appropriate analysis and time window. Parallel prerequisites can overlap; the longer barrier may govern actual start while the shorter barrier still matters to the counterfactual.

Do not add event durations or use an arbitrary percentage split to allocate an overlapping completion effect. The contractual treatment of concurrency, float, mitigation and compensation must be assessed under the actual terms and applicable law. The controls analyst should report network behaviour and alternative tests, with specialist interpretation kept explicit.

Delay, disruption, prolongation and acceleration

Delay concerns time effects; disruption concerns impaired productivity or efficiency; prolongation concerns extended time-related cost; acceleration concerns action to achieve an earlier outcome than an otherwise applicable plan. They can interact, but a delay calculation is not a disruption valuation and a cost premium is not proof that acceleration was instructed.

For disruption, compare genuinely comparable work and isolate differences in scope, labour mix, location, learning, weather, access, workfronts and acceptance. Where records support a suitable unaffected comparison, productivity comparisons can help test loss. A budget-versus-actual difference alone does not identify the causal loss or establish recoverability.

Cost traceability and duplicate recovery

Trace claimed items to quantities, resources, rates, dates, actual records and the relevant cause. Distinguish incurred, committed, forecast, estimated and already-paid amounts. Establish whether costs are incremental, fixed, redeployed, avoidable or already included in another valuation, variation or allowance.

Reconcile line items to the ledger while retaining exceptions. A paid invoice proves a transaction, not necessarily its relationship to the event. Do not price the same resource as standby, disruption and prolongation for the same period without identifying separate supported effects. Preserve offsets, mitigation, credits and the basis of each exclusion.

Neutrality, uncertainty and reproducibility

An analyst may be engaged by one party but should still identify evidence that weakens the proposed interpretation. Distinguish agreed facts, disputed facts, analytical assumptions, technical conclusions and commercial/legal questions. Explain gaps and what further evidence could change the result.

Retain input files, source references, model settings, calculation steps, adjustments and alternative runs. An AI-generated chronology or narrative can assist organisation, but verify every source, date and causal statement. Never fill missing records with plausible invented events or present a generated quotation as contemporaneous evidence.

3. Step-by-step method

1. Define the question and remit. Record event, period, affected obligation, claimed remedy, audience, contractual context and questions requiring specialist interpretation.

2. Preserve relevant records. Secure originals, versions, metadata and access/retention arrangements; document analytical copies and transformations.

3. Build an evidence index. Identify source, date, owner, reliability, related event/activity/cost item and any missing or contradictory information.

4. Establish the chronology. Separate occurrence, instruction, discovery, effect, mitigation and resolution; reconcile dates against contemporaneous records.

5. Confirm the references and time perspective. Identify baseline, update, data date, current status and whether the analysis is prospective or retrospective.

6. Validate the models. Check logic, calendars, actuals, constraints and revisions; disclose repairs and preserve unrepaired versions for comparison.

7. Trace cause and effect. Identify work prevented, resources, prerequisites, controlling/near-controlling paths, acceptance gates and alternative explanations.

8. Select and justify the method. Match the question and available evidence; identify assumptions and limits before performing schedule or productivity calculations.

9. Test events and alternatives. Analyse overlapping events, mitigation, path changes and counterfactuals without adding effects mechanically or assuming responsibility from chronology.

10. Examine cost quantum. Trace items to records, cause, quantity, rate and period; reconcile incurred/forecast/paid amounts and remove or flag duplicates and offsets.

11. Separate findings from entitlement. State technical effects and evidence status; refer notice, contract interpretation, responsibility, concurrency treatment and recoverability to authorised specialists.

12. Review and issue reproducibly. Obtain appropriate peer/specialist challenge; provide chronology, calculations, alternatives, limitations and source references, preserving the analysis version and subsequent decisions.

4. Worked Reference Project example — Community & Learning Centre

Context — a separate forensic teaching window

This is a simplified, independent event window within WBS 1.4 and WBS 1.6. It is not an additional delay or cost added to the M5 forecasts or the CR-M6-01 branch. The project remains the 24-month, 50.0m CU centre with fixed public opening. Relative day numbers below are working-day boundaries, not invented calendar dates; tasks starting at boundary 0 with duration five finish at boundary 5.

At the start of the window, approved installation requires both room access and a valid configuration. The planned model has both prerequisites available at Day 0, installation lasting five working days and tests/handover lasting three. Another readiness path finishes at Day 6. Public opening is fixed at Day 12. The supplier alleges that access provided only at Day 5 caused five days of delay and requests five days' time adjustment plus 0.090m CU.

Step A — establish facts and the missing interpretation

The signed access log supports access at Day 5. A failed configuration check and approval record support configuration readiness at Day 3, not Day 0. Installation records show start Day 5 and finish Day 10; accepted test/handover records show readiness Day 13. The other readiness path finished Day 6.

The event register records source references E01 access log, E02 configuration records, E03 installation diary, E04 acceptance/test evidence and E05 schedule exports. These are illustrative evidence IDs. The analyst checks that the records use the same work area, configuration and working-day calendar. Contractual responsibility for access, configuration and any time/money remedy remains a separate specialist question.

Step B — reproduce the simple network

Installation start = later of access availability and configuration readiness. Installation finish = start + 5; Technology readiness = finish + 3. Project readiness = later of Technology readiness and the other path's Day 6 finish.

The observed readiness movement from the planned Day 8 is five working days, while the fixed-opening miss is one working day: Day 13 versus opening Day 12. The four days between planned readiness and opening are consumed in this model; they should not disappear from the explanation. A five-day movement of readiness and a one-day miss of opening answer different questions.

Adding the isolated five-day and three-day event effects would suggest eight days of delay, which the model does not support. The barriers overlap at the prerequisites. This simple calculation demonstrates network interaction; it does not by itself determine contractual concurrency or a time entitlement.

Step C — challenge attribution with alternative tests

Removing the access event while retaining the observed Day 3 configuration condition produces readiness Day 11. Removing the configuration event while retaining Day 5 access produces Day 13. The difference between Day 13 and Day 11 is two working days in the first counterfactual, while the second changes no actual readiness date.

Those tests are informative but are not an automatic allocation of “two compensable days” to access and “three days” to configuration. Each holds particular conditions fixed and answers a defined technical question. The analyst must explain the selected method, any resource/mitigation assumptions and how the applicable contractual assessment treats overlapping causes. The result should not silently switch from the five-day isolated access effect to the two-day removal test because one favours a preferred outcome.

Step D — reconcile the monetary request

The supplier's request comprises crew standby five days × 0.010m CU = 0.050m CU; equipment five days × 0.004m CU = 0.020m CU; site overhead five days × 0.003m CU = 0.015m CU; and an administration amount of 0.005m CU. Total = 0.090m CU.

The review distinguishes 0.042m CU of evidenced candidate crew/equipment cost, 0.015m CU of recorded overhead with unresolved incremental/causal treatment, and 0.033m CU not supported as claimed: 0.020 redeployed crew, 0.008 equipment basis gap and 0.005 administration. The categories reconcile to 0.090m CU.

The 0.042m CU is not a payment recommendation or established recoverable amount. It still requires causal, contractual, rate and duplicate-payment checks. Redeployed crew cost may be genuine project expenditure elsewhere but is not supported as idle standby for those days. Recorded overhead may include cost that would have arisen anyway. The analyst seeks the necessary counterfactual and period evidence rather than accepting all five days' charges as additional loss.

Step E — issue a bounded finding

A defensible technical summary states that the available records support access at Day 5, configuration availability at Day 3 and readiness at Day 13. The simplified model reproduces five days' readiness movement and one day's miss of the fixed opening. Alternative runs demonstrate overlapping prerequisite effects, so event durations are not additive.

The cost review identifies evidence status and gaps for every claimed line and reconciles the full request. The report lists outstanding notice/contract interpretation, event responsibility, concurrency treatment, incremental overhead and duplicate-payment questions for the authorised commercial/legal specialists. It also records that the analysis is a simplified window rather than a complete whole-project forensic conclusion.

The controls lesson is to preserve access, configuration, resource redeployment and acceptance records at the time of the event. Better records can support timely mitigation and reduce disputed recollection; they do not predetermine a party's entitlement.

5. Interpretation and professional judgement

Criticality must be demonstrated in the relevant window

A current software label does not establish historical criticality. Examine the path and status at the time, then explain later path changes. The simplified example holds the other path at Day 6; if evidence showed that it actually finished Day 14, the project-readiness analysis would change and the claimed completion effect would need reassessment.

Do not confuse an analytical result with a legal conclusion

The five-day isolated effect, two-day removal-test difference and one-day opening miss describe different comparisons. None is automatically the granted time remedy. Show all relevant definitions and let the appropriate contractual analysis determine how they apply. The same separation is needed between evidenced expenditure and compensation.

Missing records require qualification, not invention

If configuration approval cannot be dated reliably, test justified alternative dates and show sensitivity. A plausible narrative is not a substitute for a contemporaneous record. State which conclusion remains supportable and what evidence would resolve the gap.

Be even-handed about mitigation and contrary evidence

Redeployment may reduce standby loss while demonstrating useful mitigation. It should be preserved even if it weakens the original request. Conversely, a proposed alternative workfront is not actual mitigation if it was unavailable or incompatible with required resources. Test feasibility from records rather than asserting that either party should simply have worked elsewhere.

6. Practice diagnostic

Use this table to challenge the quality of the approach or deliverable. The rows consolidate the most important recurring red flags and checks; they are prompts for professional judgement rather than a substitute for the method above.

7. Try-it-yourself practical exercise

Prepare a short forensic briefing for the access/configuration window.

·       1. Build the event/evidence index and chronology, identifying what each source proves and what it does not.

·       2. Reproduce the four schedule runs and distinguish readiness movement from the opening miss.

·       3. Explain why five plus three does not equal eight days of project delay in this network.

·       4. Show the two removal tests and explain why they do not automatically allocate entitlement.

·       5. Recalculate if configuration is available at Day 7 with access at Day 5; then if the other readiness path actually finishes at Day 14 in the original case.

·       6. Reconcile the 0.090m CU request to evidence categories and list additional records needed for overhead and duplicate-payment checks.

·       7. Write a neutral technical conclusion and a separate list of questions requiring commercial/legal interpretation.

·       8. Identify three records the team should improve prospectively to prevent similar evidential gaps.

What good performance looks like

The original model produces readiness Day 13: five days later than planned Day 8 and one day after opening Day 12. With configuration at Day 7, readiness becomes Day 15, seven days later than planned and three after opening. If the other path instead finishes Day 14 in the original case, actual project readiness becomes Day 14; allocation of completion effects requires examination of that path's history and causes.

The cost categories are 0.042m CU evidenced candidate cost, 0.015m CU recorded overhead requiring further analysis and 0.033m CU unsupported as claimed. The briefing does not treat any category as an agreed liability, and it retains contrary evidence and clearly labelled hypotheses.

Final challenge: “Would the same evidence and method produce the same technical account if the analysis were commissioned by the other party?”

8. Bridge to capability

The How-to teaches how records support reproducible technical cause-and-effect analysis. At APPLIED and above, capability assessment should test whether the practitioner can investigate a realistic event history, justify the method and distinguish findings from assumptions and entitlement questions. Preserved evidence, chronology, alternative-cause tests and traceable cost analysis should support a bounded conclusion. Supporting work products are illustrative and should be assessed with the practitioner’s reasoning. Sample learning questions illustrate the judgement at each recognition level; they are discussion prompts rather than validated assessment items.

Learning question: Do you understand how evidence supports an account of cause and effect?

Capability question for APPLIED and above: Can you demonstrate a balanced, reproducible analysis and distinguish supported technical findings from assumptions and entitlement questions?