Project Controls, Performance Analytics and Decisions

Introduction and context

Infrastructure projects convert approved scope into physical assets under conditions of uncertainty. Roads, bridges, rail systems, water networks and public buildings involve long supply chains, changing site conditions, regulated interfaces and substantial public or investor scrutiny. A project may appear busy while still losing control: drawings may be issued late, quantities may grow, productivity may decline, or risks may be transferred into future cost and schedule exposure.

Project controls provide the management system for detecting such conditions early and supporting disciplined decisions. They integrate the scope baseline, schedule baseline, cost baseline and risk information into a coherent performance model. Performance analytics then transform project records into indicators, trends, forecasts and decisions. Control is not merely the production of reports. It is a closed loop: establish a plan, measure actual performance, compare results with the plan, diagnose causes, decide on action, implement the action and verify its effect.

The central principle is that performance information must be linked to authorized scope and reliable measurement rules. A percentage-complete estimate without a defined work package is weak evidence; a cost variance without reference to physical progress is easily misinterpreted. Effective project controls therefore combine technical understanding, commercial awareness, data quality and professional judgement.

Learning objectives

By the end of this chapter, students should be able to:

1. Explain how scope, schedule, cost and risk baselines are integrated.
2. Apply earned value management to measure cost and schedule performance.
3. Select useful key performance indicators and design an informative dashboard.
4. Interpret trends, forecasts, uncertainty and leading indicators.
5. Explain the purpose and operation of change control.
6. Recommend evidence-based corrective actions in an infrastructure project.
7. Recognize limitations, ethical risks and governance requirements in project analytics.

1. The integrated control system

A baseline is an approved reference against which actual performance is assessed. It is not an immutable promise; authorized changes may revise it. The principal baselines are:

• Scope baseline: the approved scope statement, work breakdown structure (WBS), WBS dictionary, requirements and acceptance criteria.
• Schedule baseline: the logic-linked, resource-informed and calendar-based time plan, including activities, relationships, milestones and constraints.
• Cost baseline: the time-phased authorized budget for measured work, normally excluding management reserve.
• Risk baseline or risk model: the structured record of threats and opportunities, their causes, probabilities, impacts, owners, responses and residual exposure.

Integration is achieved through a common coding structure. A work package should have a defined deliverable, responsible owner, schedule activities, budget, measurement method and relevant risks. Control accounts aggregate work packages for managerial reporting. The WBS, cost breakdown structure, organizational breakdown structure and schedule activity codes should be cross-referenced so that a cost transaction can be connected to a physical deliverable and reporting period.

The approved budget distributed across time is the planned value (PV) profile. Actual expenditure is actual cost (AC). The value of completed work, measured using an approved earning rule, is earned value (EV). These concepts should be applied at a consistent status date and currency basis. For example, progress on a concrete work package might be earned by accepted cubic metres installed, by weighted milestones, or by a 50/50 rule for short-duration tasks. Subjective progress percentages should be supported by records such as inspection approvals, delivery tickets, survey data or test results.

A practical data flow is:

Figure: Project controls as a feedback loop from baselines through measurement and decision.

Baseline quality and status dates

A control system is only as credible as its baseline and data. Baseline quality requires clear scope, realistic quantities, logical schedule relationships, identifiable assumptions, traceable estimates and approved contingency treatment. The status date must be common to progress, cost, schedule and risk reporting. If cost is current to 30 June but progress is current to 15 June, apparent variances may be reporting artefacts.

Baseline changes should preserve history. A revised baseline should not overwrite the original plan; it should show the approved change, effective date, reason, decision authority and effect on scope, time, cost and risk. This permits accountability and lessons learned.

2. Earned value management

Earned value management (EVM) integrates scope, schedule and cost by comparing the value of work planned, the value of work actually completed and the cost incurred. The basic measures are:

The principal variances are:

where SV is schedule variance and CV is cost variance, both expressed in currency units. A positive SV means more budgeted work has been earned than planned at the status date; a positive CV means the earned work cost less than its budgeted value. Schedule variance in EVM is not a direct number of days and should not replace critical-path schedule analysis.

Efficiency indices are:

SPI is the schedule performance index and CPI is the cost performance index. Values greater than 1.00 indicate favourable performance relative to the relevant baseline; values below 1.00 indicate unfavourable performance. These ratios are dimensionless and are meaningful only when PV, EV and AC use comparable scope, time period and currency conventions.

A common forecast, assuming current cost efficiency continues, is:

If the remaining work is re-estimated independently, the more general identity is:

Forecast variance is:

A simple estimate of remaining cost using current cost performance is:

These forecasts assume that the selected performance pattern is relevant to future work. If the past variance was a one-time event and the remaining work will follow the original plan, a different forecast may be justified. Forecast formulas are decision aids, not substitutes for a bottom-up estimate.

Worked example: a water-treatment package

At the status date, a mechanical installation package has BAC = $2,400,000. The baseline planned value is PV = $1,500,000, the approved value of work completed is EV = $1,320,000, and actual cost is AC = $1,440,000.

Therefore:

The package has earned 88% of the planned value by the status date and is obtaining about $0.92 of budgeted value per $1.00 of actual cost. If this cost efficiency persists:

Thus, the forecast overrun is approximately $218,182, because VAC = $2,400,000 − $2,618,182. The project team should not immediately conclude that acceleration is required. It should investigate causes: late equipment, rework, low installation productivity, design changes, payment timing or incorrect progress measurement. Corrective action must address the cause rather than merely improve the appearance of the indicator.

EVM limitations

EVM can conceal important realities. A package may have a favourable CPI because expensive materials were not yet purchased, while future procurement exposure remains. A favourable SPI may result from earning value early on non-critical work while a critical-path activity is late. Front-loaded budgets and inappropriate earning rules can distort results. EVM also treats approved scope as the reference; unapproved scope growth should be visible through change control rather than hidden in progress claims. Consequently, EVM should be read with critical-path analysis, quantities and productivity, commitments, cash flow, risk exposure and quality information.

3. KPIs, dashboards and performance analytics

A key performance indicator (KPI) is a defined measure used to monitor an important objective. A useful KPI has a clear purpose, owner, formula, data source, reporting frequency, threshold and action rule. Indicators should be balanced across outputs, drivers and constraints.

Leading indicators signal conditions likely to affect future results: design review backlog, procurement lead-time slippage, unresolved requests for information, declining labour productivity, open high-risk permits and inadequate staffing. Lagging indicators describe outcomes: final cost, achieved milestones, defects and incidents. A mature dashboard combines both.

A dashboard should answer three questions: What is happening? Why is it happening? What decision is required? It should show the status date, baseline version, data completeness, current value, trend, threshold, responsible owner and agreed action. Traffic-light colours should have explicit rules; for example, amber may mean a forecast exceeds an internal tolerance and requires management attention, while red may mean a contractual milestone or safety threshold is threatened. Colours without definitions create false certainty.

Trend analysis and forecasting

Trend analysis compares observations over time rather than relying on a single result. A simple linear trend model is:

where y_t is the observed measure at time t, a is the intercept, b is the average change per time period, and ε_t is the unexplained error. The model is useful for approximately linear, reasonably consistent data; it should not be extrapolated through a known change in work conditions without reassessment.

For a sequence of values, a moving average can reduce short-term noise:

Here, k is the number of periods included. A moving average is descriptive and may lag turning points. Forecasts should include assumptions and, where possible, a range rather than a single supposedly precise number.

Risk-adjusted forecasting can be expressed conceptually through expected monetary value:

where p_i is the probability of risk event i and I_i is its monetary impact, with threats represented as negative and opportunities as positive if that convention is adopted. EMV is an average over repeated comparable situations; it does not predict the exact outcome of one project and should not be used alone for contingency decisions.

4. Risk-informed decisions

Risk management and project controls must interact continuously. The risk register should identify affected WBS elements, schedule activities, cost accounts, assumptions, triggers, response owners and residual risk. A risk that occurs becomes an issue and should enter issue and change processes. A quantitative schedule or cost risk analysis may use probability distributions and simulation, but the quality of outputs depends on assumptions, correlations and data.

Decision quality improves when alternatives are assessed against explicit criteria such as safety, compliance, time, whole-life cost, constructability, carbon, stakeholder impact and residual risk. A decision record should state the evidence considered, assumptions, options rejected, authority, decision date and review trigger. This is especially important where a short-term saving creates maintenance, safety or resilience costs later.

5. Change control

Change control protects the integrity of the approved plan while allowing legitimate adaptation. Typical changes include design development, regulatory requirements, unforeseen ground conditions, client scope additions, compensation events and approved risk responses.

A robust change process includes:

1. Record the proposed change and its origin.
2. Describe the affected requirement, drawing, specification or contract clause.
3. Analyse scope, quantities, cost, schedule, quality, safety, environment, interfaces and risk.
4. Identify alternatives and the consequence of doing nothing.
5. Obtain review and authorization at the correct authority level.
6. Update the baseline, budget, schedule, risk register, forecasts and communications.
7. Track implementation and verify the outcome.

The distinction between trend, issue, risk and change is useful. A trend is an emerging movement that may become material; an issue is a current problem; a risk is an uncertain future event; a change is an alteration to approved scope, baseline or contract terms. Poor projects often use informal instructions and later attempt to reconstruct authorization, causing disputes and unreliable forecasts.

Change control must not become bureaucracy that prevents necessary action. Emergency safety actions may proceed under delegated authority, but the rationale and retrospective authorization should be documented. Contractual notice periods, records requirements and delegated powers must be respected.

6. Corrective action and decision logic

Corrective action should follow diagnosis. A practical sequence is:

• Detect: identify a material variance, threshold breach or adverse trend.
• Validate: check data, status date, coding, measurement rules and baseline version.
• Diagnose: use root-cause analysis, process mapping, site observation and responsible-owner interviews.
• Evaluate: compare options, including their effects on safety, quality, cost, time, risk and stakeholders.
• Authorize: select an action within delegated authority or escalate it.
• Implement: assign owner, resources, due date and measurable success criteria.
• Verify: determine whether the underlying condition improved and capture learning.

Possible actions include resequencing work, removing design constraints, changing procurement strategy, increasing supervision, improving crew balance, negotiating an approved commercial change, accepting a low-value variance, or revising the forecast. “Spend more to recover time” is not automatically rational: acceleration may increase congestion, rework and safety exposure. Decision-makers should consider marginal effects and the critical path.

Illustrative case: urban light-rail corridor

Background

An illustrative light-rail corridor project includes track formation, drainage, utility diversions and station works. The project baseline budget is $86 million, with a contractual completion milestone in 18 months. At month 9, the dashboard reports CPI = 0.94, SPI = 0.91 and two months of negative critical-path float. The contractor proposes adding a second shift.

Problem

The headline indicators suggest both cost and schedule underperformance, but the project board must determine whether additional shifts are the best response. The data also show that utility approvals are late, station excavation productivity has fallen, and several change requests remain unpriced.

Analysis

The controls team first validates the numbers. It finds that drainage progress is measured by installed length, while some excavated but untested sections have been credited prematurely. After correction, drainage EV falls by $0.4 million. The cost report includes invoices for long-lead rail already delivered, whereas the schedule effect of a delayed signalling design is not represented in EVM. Critical-path analysis identifies signalling design approval and utility diversion—not track formation—as the controlling sequence.

A cause analysis identifies three principal drivers: incomplete utility information, delayed design decisions and restricted work fronts. A second shift on track formation would increase cost but would not remove the controlling constraints. Risk review also identifies a high probability of night-work community restrictions.

Approach and solution

The board rejects immediate blanket acceleration and approves a staged response. First, the client and designer establish a ten-day utility decision cell with named authorities. Second, the contractor resequences unaffected drainage and station enabling works. Third, a limited night shift is piloted only where permits, lighting, noise controls and safe access are confirmed. Fourth, unpriced changes are separated from performance variance and processed through formal change control. The baseline is updated only for approved changes. The dashboard adds design approval age, available work-front percentage, corrected progress and critical-path float.

Outcome

After six weeks, utility decisions are current, available work fronts increase, and critical-path float stabilizes. The pilot shift improves output in one zone but is discontinued in another because congestion causes rework. The forecast remains above the original budget because an approved utility change is added transparently, while the expected completion date returns within the contractual tolerance.

Learning points

The case demonstrates that indicators are signals, not explanations. Data validation can change the diagnosis; EVM must be combined with network schedule logic; approved change must be distinguished from contractor performance; and corrective action should target constraints. It also demonstrates why a dashboard should include leading indicators and why safety and community effects belong in acceleration decisions.

7. Governance, ethics and technology

Project controls involve power: reported performance can affect payment, reputation, claims and employment. Ethical practice requires honest progress measurement, transparent assumptions, protection against retaliation for bad news, and separation of evidence from desired outcomes. Deliberately delaying cost recognition, inflating progress, hiding forecast overruns or splitting changes to avoid approval thresholds undermines governance and may breach contractual or legal duties.

Data governance should define ownership, access rights, version control, retention, audit trails and correction procedures. Personal data, commercially sensitive rates and security-related infrastructure information require appropriate protection. A dashboard should not imply more accuracy than the source data supports.

Digital delivery platforms, model-based quantities, field applications, sensors, drones and common data environments can improve timeliness and traceability. Artificial intelligence may assist with anomaly detection, document classification, delay-risk screening and forecast generation. However, AI outputs depend on representative, correctly labelled data and can reproduce historical bias or generate plausible but unsupported explanations. Human review remains essential for contractual, safety and public-interest decisions. AI-generated recommendations should be traceable to data, assumptions and model version, with a clear escalation path.

8. Best-practice implementation

An effective project controls implementation should:

• establish a clear WBS and coding dictionary before detailed reporting;
• define progress measurement rules for every significant work package;
• align schedule, cost, procurement and risk status dates;
• maintain a controlled baseline and an auditable change log;
• reconcile committed cost, actual cost, accruals and forecast cost;
• combine EVM with critical-path, quantity, productivity, quality and safety analysis;
• use leading and lagging indicators with explicit thresholds and owners;
• report uncertainty and assumptions, not only point forecasts;
• focus meetings on decisions and actions rather than presentation of data;
• close the feedback loop by verifying corrective-action effectiveness;
• preserve records for claims, audit, handover and lessons learned.

Chapter Summary

Project controls integrate approved scope, schedule, cost and risk information so that infrastructure teams can measure performance and make timely decisions. The common EVM measures are PV, EV and AC; SV and CV show value-based variances, while SPI and CPI show relative schedule and cost efficiency. EAC, ETC and VAC support forecasting, but their assumptions must be tested.

Reliable analytics depends on consistent status dates, valid progress rules, traceable data and preserved baseline history. Dashboards are most useful when they combine leading and lagging indicators, show trends and thresholds, identify owners and specify decisions. EVM should be complemented by critical-path analysis, quantities, productivity, quality, safety, commitments and risk exposure.

Change control preserves accountability while allowing authorized adaptation. Corrective action requires validation, root-cause diagnosis, option evaluation, authorization, implementation and verification. Ethical governance demands truthful reporting, auditability, protection of sensitive information and responsible use of digital and AI tools. The professional objective is not to make indicators look favourable, but to convert reliable evidence into safer, more timely and more economically sound project outcomes.

Key Terms

Baseline: An approved reference plan used for performance comparison.

Control account: A management control point where scope, schedule and cost responsibility are integrated.

Cost performance index (CPI): Earned value divided by actual cost.

Critical path: The longest controlling path through a schedule network, determining minimum planned duration under stated logic.

Earned value (EV): Budgeted value of work actually completed.

Estimate at completion (EAC): Forecast total cost of the project or control account at completion.

Estimate to complete (ETC): Forecast cost required to complete remaining work.

Forecast: A reasoned projection of future cost, time, risk or performance.

Key performance indicator (KPI): A defined measure used to monitor an important objective.

Leading indicator: A measure that signals conditions likely to affect future performance.

Planned value (PV): Budgeted value of work scheduled by a specified status date.

Progress measurement rule: An approved method for determining how work earns value.

Risk register: A controlled record of risks, causes, probabilities, impacts, owners and responses.

Schedule performance index (SPI): Earned value divided by planned value.

Status date: The common date to which project performance information is reported.

Trend: An emerging direction or pattern in observations over time.

Variance at completion (VAC): Budget at completion minus estimate at completion.

Variance analysis: Investigation of the magnitude, cause and significance of deviation from a baseline.

Work breakdown structure (WBS): A hierarchical decomposition of the total approved project scope.

Review Questions

Short Answer Questions

1. What is the purpose of integrating scope, schedule, cost and risk baselines?
2. Define PV, EV and AC in earned value management.
3. Why is EVM schedule variance not equivalent to a delay expressed in days?
4. Distinguish a risk, an issue, a trend and a change.
5. Give two examples of leading indicators in an infrastructure project.

Long Answer Questions

1. Explain how a common coding structure supports integrated project controls.
2. Describe the EVM formulas for SV, CV, SPI, CPI and EAC, including their assumptions and limitations.
3. Discuss the design features of an effective infrastructure project dashboard.
4. Explain a complete change-control process and its relationship to baseline governance.
5. Evaluate why corrective action should be based on root-cause analysis rather than headline KPI results alone.

Analytical / Application Questions

1. A package has BAC = $5,000,000, PV = $3,000,000, EV = $2,700,000 and AC = $3,000,000. Calculate SV, CV, SPI, CPI and an EAC based on current cost efficiency. Interpret the results and identify two further investigations required.
2. A project has a favourable CPI but its critical-path float is falling rapidly. Explain how this situation could occur and recommend a dashboard and decision response.
3. A contractor proposes a costly second shift after a project reports SPI = 0.85. Develop an evidence-based decision process that tests whether acceleration addresses the actual controlling constraint, including safety, risk, change and commercial considerations.

Multiple Choice Questions with Answers

Question 1

Which measure represents the budgeted value of work actually completed by the status date?

A. PV

B. EV

C. AC

D. BAC

Correct Answer: B. EV

Question 2

If EV is $900,000 and AC is $1,000,000, what is the CPI?

A. 0.90

B. 1.11

C. $100,000

D. $1,900,000

Correct Answer: A. 0.90

Question 3

A schedule performance index below 1.00 generally indicates that the project has:

A. Spent less than planned

B. Earned less budgeted work than planned

C. Completed all critical activities

D. Exceeded the approved budget at completion

Correct Answer: B. Earned less budgeted work than planned

Question 4

Which statement about EVM schedule variance is correct?

A. It directly states the number of days late

B. It is expressed as a value-based comparison and should be complemented by schedule analysis

C. It measures only safety performance

D. It is calculated as AC minus EV

Correct Answer: B. It is expressed as a value-based comparison and should be complemented by schedule analysis

Question 5

If BAC is $2,000,000 and CPI is 0.80, the simple EAC forecast based on continuing cost efficiency is:

A. $1,600,000

B. $2,000,000

C. $2,500,000

D. $3,000,000

Correct Answer: C. $2,500,000

Question 6

Which is a leading indicator?

A. Final project cost

B. Completed defects at handover

C. Design approval backlog

D. Final completion date

Correct Answer: C. Design approval backlog

Question 7

What is the primary purpose of a controlled baseline change log?

A. To conceal earlier forecasts

B. To preserve authorization, history and accountability for changes

C. To eliminate all project changes

D. To replace the risk register

Correct Answer: B. To preserve authorization, history and accountability for changes

Question 8

Before taking corrective action on an adverse KPI, the first professional step should normally be to:

A. Increase the budget immediately

B. Validate the data, status date and measurement rules

C. Delete the KPI

D. Accelerate every activity

Correct Answer: B. Validate the data, status date and measurement rules

Question 9

Which formula represents cost variance in EVM?

A. PV − EV

B. EV − AC

C. AC − BAC

D. BAC − PV

Correct Answer: B. EV − AC

Question 10

Why should AI-generated project-control recommendations receive human review?

A. AI outputs are always legally binding

B. AI cannot process any project data

C. Outputs depend on data and assumptions and may contain unsupported or biased conclusions

D. Human review makes baselines unnecessary

Correct Answer: C. Outputs depend on data and assumptions and may contain unsupported or biased conclusions