9.5 Requirements Elicitation, Plan Implementation & Plan Validation
Key Takeaways
- AACE RP 11R-88 places Requirements Elicitation and Analysis first in the Planning process because every later estimate, schedule, and control account inherits whatever the requirements got wrong.
- Requirements must be verifiable: a stated need becomes a requirement only when it carries a measurable acceptance criterion, a source, and a traceable link to a WBS deliverable.
- Plan implementation has two halves: project implementation authorises the work, and project control implementation stands up the measurement system before the first status report is due.
- Plan validation is a gate, not a review meeting: the plan is validated only when scope, estimate, and schedule are mutually consistent, resource-feasible, and constructible before the baseline is frozen.
- Schedule health checks such as the DCMA 14-point assessment flag structural defects including open ends, excessive lags, hard constraints, negative float, and high-duration activities before the baseline is accepted.
9.5 Requirements Elicitation, Plan Implementation & Plan Validation
Quick Summary: AACE RP 11R-88's Planning process begins with Requirements Elicitation and Analysis and its Plan Implementation process ends with Plan Validation. Between them sit Project Implementation and Project Control Implementation. These four steps are the connective tissue between Chapter 9's WBS and schedule work and Chapter 11's earned value reporting: they turn an approved scope into an authorised, measurable, validated baseline. Projects that skip them do not fail at handover; they fail in month three, when the first variance report cannot be trusted because nobody agreed what was being measured.
1. Requirements Elicitation and Analysis
Why It Is First
A requirement error is the cheapest defect to fix and the most expensive to leave. It propagates into the WBS, the estimate, the schedule, the procurement packages, and the acceptance criteria, and it is usually discovered at commissioning. Because it precedes scope definition, it sits at the extreme left of the cost influence curve, where leverage is maximal.
Elicitation Techniques
| Technique | Best for | Watch out for |
|---|---|---|
| Structured interviews | Individual expert knowledge, operations practice | Interviewee describes the current plant, not the future need |
| Facilitated workshops | Cross-discipline conflicts surfaced in one room | Dominant personalities capture the outcome |
| Document analysis | Regulatory obligations, permits, standards, existing as-builts | As-builts may not match the plant — see configuration management |
| Observation / walkdown | Tacit operating practice nobody writes down | Observer effect; unsafe access |
| Questionnaires and surveys | Large or dispersed stakeholder groups | Low response rate biases the sample |
| Delphi | Contested technical judgements needing bias control | Slow; needs a disciplined facilitator |
| Prototyping / 3D review | Layout, access, maintainability | Cost of the model versus the value of the finding |
Turning a Need Into a Requirement
A statement becomes a usable requirement only when it is:
- Unambiguous — one interpretation. "Adequate lighting" is a wish; "300 lux maintained at the working plane" is a requirement.
- Verifiable — it carries an acceptance test. If nobody can say how it will be proven at handover, it cannot be estimated, scheduled, or accepted.
- Traceable — it has a source (a person, a regulation, a standard) and a destination (a WBS deliverable).
- Feasible and necessary — achievable within the project's constraints, and required rather than merely desired.
- Prioritised — classified, for example must-have / should-have / could-have, so that value engineering and scope trade-offs have a rule to follow.
The Requirements Traceability Matrix
The traceability matrix is the artefact that makes requirements auditable. Each row carries: requirement ID, description, source, priority, the WBS element that delivers it, the verification method, and the current status.
REQ-ID | Requirement | Source | Pri | WBS | Verified by
--------+-----------------------------------+---------------+------+---------+------------------
R-014 | Firewater ring main 1,500 gpm | NFPA 24 / | Must | 3.2.1 | Witnessed flow
| at 100 psi residual | Insurer | | | test at handover
R-027 | Control room noise <= 60 dB(A) | Occupational | Must | 4.1.3 | Sound survey,
| | health std | | | 3 locations
R-051 | Spare 20% conduit fill for future | Operations | Could| 5.4.2 | Drawing review
A requirement with no WBS element is unfunded and unscheduled. A WBS element with no requirement is scope gold-plating. Both are found by reading the matrix down each column.
2. Scope and Execution Strategy Development
Requirements analysis feeds directly into the execution strategy: how the work will be split, contracted, and sequenced. The strategy decisions a cost technician must be able to recognise are:
- Delivery method — design-bid-build, design-build, EPC, EPCM, construction management at risk, integrated project delivery.
- Contracting strategy — one lump-sum wrap versus multiple discipline packages; the risk allocation consequences are developed in Chapter 12.
- Work packaging and phasing — how the WBS is cut into procurable, schedulable units, and whether the project is fast-tracked.
- Self-perform versus subcontract — drives the direct/indirect split and the labour burden exposure.
- Site logistics and modularisation — stick-build versus modular fabrication changes the schedule shape, the risk profile, and the location factor that applies.
Execution strategy is not an administrative preamble: it determines the estimate basis, so it must be settled before the control estimate is prepared.
3. Plan Implementation, Part One — Project Implementation
Project implementation converts an approved plan into authorised, funded work.
- Work authorisation. A formal, documented release of scope, budget, and schedule dates to a named Control Account Manager (CAM). No CAM may charge to a control account before the authorisation exists. This is the mechanism that prevents the two classic accounting failures: charging work that was never approved, and performing approved work with no budget to charge to.
- Budget distribution. The Budget at Completion is decomposed downward: BAC → control accounts → work packages and planning packages, with Undistributed Budget for authorised scope not yet allocated, and Management Reserve held outside the Performance Measurement Baseline.
- Charge-number and code-of-accounts setup. Cost codes must mirror the WBS and OBS exactly, or actual cost will never reconcile to earned value.
- Procurement release. Purchase orders and subcontracts are issued against the authorised packages; commitments begin to accrue.
- Mobilisation. Site establishment, temporary facilities, craft ramp-up per the resource histogram.
4. Plan Implementation, Part Two — Project Control Implementation
The control system must be operating before the first reporting period closes. Standing it up late is the most common reason a project's first three status reports are useless.
| Element | What must exist before period 1 closes |
|---|---|
| Performance Measurement Baseline | Time-phased, resource-loaded, formally approved; the S-curve exists |
| Progress measurement rules | Each work package assigned a technique — units completed, incremental milestones, 0/100, 50/50, percent complete, level of effort — and written into the WBS dictionary |
| Data date and reporting calendar | Cost cut-off and schedule data date aligned; if payroll closes Friday and the schedule is statused Wednesday, every variance is noise |
| Cost collection | Charge codes live; committed, incurred, and paid distinguished; accruals defined for work done but not invoiced |
| Variance thresholds | The values that trigger a written variance analysis report, by dollar and by percentage |
| Change control | Trend register and change order log open; the board is named and meets on a schedule |
| Reporting formats | The report a CAM must produce, and the one the sponsor will read, both agreed and templated |
[!WARNING] Misaligned cut-off dates are the silent killer. If actual cost (AC) is collected to month-end but earned value (EV) is statused mid-month, the resulting CPI is measuring a calendar artefact, not performance. Align the data dates before period 1, not after the first argument.
5. Plan Validation — the Gate Before the Baseline Freezes
Plan validation asks one question: is this plan capable of being executed as written? It is a gate, not a review meeting, and it occurs before the baseline is frozen.
The Four Consistency Tests
- Scope ↔ estimate. Every WBS element is priced; every priced item maps to a WBS element; the Basis of Estimate's exclusions, assumptions, and allowances are consistent with the requirements matrix.
- Scope ↔ schedule. Every work package appears in the network; no activity exists without a work package.
- Estimate ↔ schedule. The estimate's labour hours match the schedule's resource loading. A 240,000-hour estimate against a schedule whose histogram integrates to 180,000 hours is not a plan; it is two documents.
- Schedule ↔ resources ↔ site. The peak craft density implied by the histogram must fit the physical workface. A peak of 900 craft on a site that can safely hold 550 is an invalid plan regardless of how the network calculates.
Schedule Health Check
Structural defects are found by rule, not by reading. The widely used DCMA 14-point assessment checks, among others:
- Logic — activities with no predecessor or no successor (open ends); target ≤ 5%.
- Leads (negative lags) — target zero.
- Lags — target ≤ 5% of relationships.
- Relationship types — finish-to-start should dominate (≥ 90%).
- Hard constraints — target ≤ 5%; "must finish on" constraints mask float.
- High float — activities with total float > 44 days; target ≤ 5%.
- Negative float — target zero in a baseline.
- High duration — activities longer than 44 working days; target ≤ 5%.
- Invalid dates — no actual dates after the data date, no forecast dates before it.
- Resources — every non-LOE activity carries resources or cost.
- Missed tasks, critical path test, CPLI, BEI — behavioural checks applied once the schedule is being statused.
Estimate Validation
- Independent review by estimators not on the project team.
- Benchmarking against normalised historical unit rates and against industry metrics such as cost per square foot, per installed tonne, or per kW.
- Class conformance — does the degree of project definition actually support the class being claimed (see Section 7.2)?
- Risk and contingency reconciliation — does the contingency reflect the risk register, and is the confidence level (P50, P80) the one the governance policy requires?
Constructability and Operability Review
A cross-functional walkthrough by construction and operations staff, asking: can it be built in this sequence, with this access, by crews of this size, and can it then be started up, operated, isolated, and maintained? Findings at this gate are drawing changes. The same findings six months later are rework.
Integrated Baseline Review (IBR)
On EVM-managed projects the IBR is the formal joint owner/contractor validation that the PMB is realistic, that technical scope, schedule, and budget are consistent, that risks are identified, and that the management processes are in place. It is a management review, not an audit: its output is a list of agreed actions, and only when they are closed is the baseline accepted.
[!TIP] The exam distinction to hold: validation asks "is the plan executable?" Verification asks "was the work built to the plan?" Validation happens before the baseline; verification happens during and after execution.
6. Worked Scenario: A Baseline That Should Not Have Been Frozen
A mid-size EPC project submits its baseline for acceptance. The control team applies the validation gate.
- Scope ↔ estimate: 3 of 41 work packages are in the estimate but not the WBS dictionary. Rejected pending correction.
- Estimate ↔ schedule: estimate carries 186,000 direct craft hours; the resource-loaded schedule integrates to 154,000. A 32,000-hour gap — 17% — means either the estimate is padded or the schedule is under-resourced. Investigation shows the schedule omitted the scaffolding subcontract entirely.
- Schedule health: 11% of activities are open-ended, there are 6 "must finish on" constraints, and 4 activities carry negative float of −9 days before any progress has been recorded. A baseline cannot start with negative float.
- Resource feasibility: peak craft 740 against a site access study limit of 560.
- Control system readiness: progress measurement technique is unassigned for 12 work packages; cost cut-off is the last calendar day of the month while the schedule data date is the 25th.
Outcome: the baseline is not accepted. Four weeks of rework to the schedule and estimate follow. Had it been accepted, the project would have reported a plausible CPI and SPI for two quarters while being structurally incapable of delivering — the exact failure mode that plan validation exists to prevent.
7. Exam Watch: High-Yield Traps & Rules of Thumb
[!CAUTION] A requirement without a verification method is not a requirement. Options describing "adequate", "state of the art", or "industry standard" performance are wishes. The verifiable, measurable statement is always the right answer.
[!WARNING] Work authorisation precedes charging. No control account may accumulate actual cost before scope, budget, and dates have been formally released to a named CAM. An option that has work starting "to protect the schedule" while authorisation is pending is describing a control failure.
[!TIP] Validation is before the baseline; verification is after the work. Validation = "can this plan be executed?" Verification = "was it built as planned?"
[!CAUTION] A baseline may not contain negative float. Negative float in a baseline means the plan already fails its own constraint. Negative float during execution is a live warning that requires compression. Do not confuse the two.
[!WARNING] Align the cost cut-off with the schedule data date. Mismatched dates make CPI and SPI meaningless, and no amount of later analysis recovers the lost periods.
During requirements analysis for a pharmaceutical fit-out, an operations manager states that the clean corridor must have 'adequate air changes to keep the space comfortable and compliant.' A cost technician flags the statement as unusable. What specific quality is missing, and what is the cost consequence of carrying it forward?
A control team is validating a submitted baseline. The approved estimate carries 186,000 direct craft hours, while the resource-loaded CPM schedule integrates to 154,000 hours for the same scope. What does this discrepancy signify, and what is the correct action?
A submitted baseline schedule contains four activities carrying negative float of -9 days before any progress has been recorded, 11% of activities with no predecessor or no successor, and six 'must finish on' hard constraints. Under standard plan validation practice, how should the control team treat this submission?