Monitoring, Control and Resource Analytics

Introduction

Resource loading and leveling do not end when a baseline is approved. Once work begins, managers must compare planned resource demand with actual effort, completed value, remaining work, capacity, and emerging constraints. Monitoring transforms a static resource plan into a controlled management system. The purpose is not merely to report whether a project is late or over budget, but to determine why performance differs from the baseline, whether the remaining plan is feasible, and what corrective action is justified.

This chapter integrates earned value management (EVM), earned schedule (ES), resource utilization, throughput, variance analysis, and forecasting. It also considers the limitations of metrics and the governance required to prevent misleading performance reporting.

Learning Objectives

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

1. Distinguish actual effort, remaining work, planned value, earned value, and actual cost.
2. Calculate and interpret cost and schedule performance indicators.
3. Apply earned schedule to improve time-based forecasting.
4. Analyze utilization, throughput, capacity, and resource variance.
5. Estimate estimate-at-completion and time-to-completion under different assumptions.
6. Construct a resource-feasible reforecast and recommend control actions.
7. Design reliable dashboard indicators while recognizing metric limitations.

1. The Control Logic of Resource Analytics

A project control system links four elements:

• Baseline: the approved time-phased plan for scope, schedule, cost, and resources.
• Actuals: recorded labor hours, costs, progress, completed deliverables, and dates.
• Forecast: an estimate of future effort, cost, duration, and resource demand.
• Action: an approved intervention, such as resequencing, staffing, scope clarification, or corrective training.

The basic control cycle is:

Plan → Measure → Analyze variance → Reforecast → Decide → Implement → Re-measure.

A useful distinction is between performance measurement and performance management. A metric describes a condition; management interprets its cause and selects an intervention. For example, a utilization rate of 96% may indicate productive deployment, excessive workload, hidden overtime, or inaccurate capacity assumptions. The number alone does not identify the correct action.

Resource control should operate at a level that is meaningful for decisions. A portfolio dashboard may use aggregated labor cost and milestone indicators, whereas a work-package manager may need task-level remaining hours, skill mix, and queue data.

2. Fundamental Measures and Data Integrity

2.1 Planned value, earned value, and actual cost

Earned value management uses three core quantities measured at a status date:

PV and EV may be expressed in budget-hours when the objective is resource analysis. AC is normally monetary, although actual effort hours can be tracked separately. Mixing units without a defined conversion is invalid.

The total approved budget for the performance measurement baseline is commonly called Budget at Completion (BAC). The baseline must contain a clear rule for earning value, such as 0/100, 50/50, weighted milestones, or objectively measured physical progress. Subjective progress claims can produce unreliable EV.

2.2 Variance measures

The principal EVM variances are:

where CV is cost variance and SV is schedule variance. In monetary EVM, a positive CV means that the value earned exceeds actual cost. A positive SV means that more budgeted work has been accomplished than was scheduled. These interpretations assume consistent valuation and a valid baseline.

A schedule variance expressed in currency is not elapsed time. It is a difference in budgeted value. Consequently, a project may have SV = 0 at completion even if it finished late. This limitation motivates earned schedule analysis.

2.3 Performance indices

CPI is cost efficiency; SPI is schedule efficiency in value terms. A value below 1.00 indicates unfavorable performance, while a value above 1.00 indicates favorable performance relative to the baseline. The indices are ratios and have no currency unit.

These measures should not be interpreted mechanically. A CPI of 0.90 may result from a one-time equipment purchase, inaccurate accruals, or a persistent productivity problem. Likewise, SPI can be distorted by front-loaded or back-loaded budgets and by the choice of progress measurement method.

3. Actual Effort and Remaining Work

3.1 Actual effort

Actual effort is the labor time recorded against project work. It should be distinguished from elapsed duration, attendance, and cost. Eight hours of work by one analyst and four hours each by two analysts are both eight labor-hours, but they may have different productivity and coordination consequences.

Time-phased actual effort should identify:

• work package and activity;
• resource role or skill;
• date or accounting period;
• productive effort, support effort, and rework;
• approved overtime or premium time;
• non-project interruptions and waiting time.

Poor time coding hides the causes of variance. A single generic project code may show total effort but cannot reveal whether work was spent on production, defect correction, meetings, or avoidable delay.

3.2 Remaining work and estimate to complete

Remaining work is the effort required to complete the defined remaining scope under stated assumptions. It is not simply the original estimate minus actual effort, because actual productivity, scope changes, defects, and learning may alter the requirement.

The current estimate of future cost is called Estimate to Complete (ETC). If labor hours are the controlling resource, an analogous estimate to complete can be expressed in hours. A credible ETC should be bottom-up where uncertainty is material: remaining tasks are estimated using current productivity, resource availability, dependencies, and risk exposure.

A common forecasting error is to treat the original baseline as an objective fact. The baseline is a decision reference, not a guarantee. A reforecast should preserve traceability by showing baseline, actual-to-date, approved changes, and current remaining estimates separately.

4. Forecasting Cost and Completion

4.1 Estimate at completion

The simplest forecast is:

where EAC is estimate at completion. This is the preferred form when ETC has been independently reassessed from the remaining work.

When the original budget is considered reliable and current cost variance is expected to be atypical, a commonly used statistical forecast is:

When current cost efficiency is expected to continue, the forecast is:

When both cost and schedule performance are expected to influence remaining work, one frequently used approximation is:

The final expression is a heuristic, not a universal law. It assumes that observed cost and schedule efficiencies are relevant to future work and that CPI and SPI are sufficiently stable. It should not replace a reasoned ETC when scope, resource mix, or operating conditions have changed.

The variance at completion is:

A negative VAC indicates an expected overrun against the approved budget.

4.2 Worked example: EVM forecast

Assume a work package has BAC = $200,000. At the status date, PV = $120,000, EV = $100,000, and AC = $125,000.

Therefore:

If cost inefficiency is expected to continue but schedule effects are not separately modeled:

Thus, VAC = $200,000 − $250,000 = −$50,000. This forecast is conditional on the assumption that the current CPI will remain representative. A bottom-up ETC could produce a different and more defensible result.

5. Earned Schedule and Time Forecasting

5.1 Rationale

Earned schedule converts earned value into time by asking: At what point in the baseline schedule should the current EV have been achieved? Let AT denote actual time elapsed at the status date. Let C be the last completed time increment for which PV does not exceed EV, and let PV_C and PV_{C+1} be planned values at consecutive time increments.

A linear interpolation estimate of earned schedule is:

ES and AT must use the same time unit, such as weeks. The formula assumes that planned value changes approximately linearly between the two reference points. If the planned-value curve is discontinuous or highly nonlinear, a more detailed interpolation method is required.

The earned schedule performance index is:

A time-based forecast under the assumption that current time efficiency continues is:

where PD is the original planned duration and EAC_t is the forecast duration. The notation EAC_t is used here to distinguish time completion from cost completion. These forecasts are approximate and should be reconciled with network logic, remaining critical-path work, and resource constraints.

5.2 Worked example: earned schedule

Suppose a project is at week AT = 10. The last completed baseline week with PV not exceeding EV is C = 8. PV_8 = $80,000, PV_9 = $100,000, and EV = $90,000.

Therefore:

If the planned duration is 20 weeks and current time efficiency continues:

This estimate should be treated as an early warning, not as a substitute for a resource-feasible schedule forecast.

6. Utilization, Capacity, and Throughput

6.1 Resource utilization

Utilization compares productive or assigned effort with available capacity. A clearly defined form is:

Net available capacity should account for holidays, approved leave, training, administrative duties, and other committed work. If “assigned effort” is used instead of productive effort, the metric measures loading rather than actual productivity. The denominator and numerator must therefore be named explicitly.

For a team:

High utilization is not always desirable. Sustained utilization near 100% leaves no capacity for variation, support, defects, or urgent work. Excessive utilization can increase queues, fatigue, and rework.

6.2 Throughput

Throughput is the rate at which acceptable units of work are completed. For a period:

Units must be comparable: tested features per week, approved drawings per month, or closed service-transition work packages per sprint. Counting activity rather than accepted output creates a misleading measure.

Little’s Law provides a useful queueing relationship under relatively stable conditions:

WIP is work in progress, throughput is output per unit time, and lead time is the average time from work entry to completion. The relationship assumes a stable system over the observation period. It indicates that increasing work in progress does not automatically increase throughput; it may instead increase lead time and congestion.

6.3 Capacity variance

A simple capacity variance compares planned demand with available capacity:

A positive result indicates a shortage when both terms use the same resource and time units. Capacity variance should be calculated by skill, location, time period, and sometimes shift—not only as a project total. A project may have sufficient total labor-hours but lack the specialist capacity required in a particular week.

7. Resource Variance Diagnosis

A useful variance investigation separates symptoms from causes. Common categories include:

Variance analysis should use thresholds. For example, a manager may require investigation when CPI falls below 0.90 for two consecutive reporting periods, when a critical skill has more than 40 hours of uncovered demand, or when forecast completion exceeds a contractual milestone. Thresholds must be tailored to risk, project size, and organizational tolerance.

8. Dashboard Design for Decision-Making

A dashboard should present indicators that lead to decisions, not merely display available data. A practical dashboard may contain:

• scope status and approved change exposure;
• PV, EV, AC, CPI, SPI, and trend lines;
• actual effort versus planned effort;
• remaining work and ETC by work package;
• forecast completion date and critical-path risk;
• utilization by role and period;
• throughput, WIP, and aging work;
• capacity gaps and overloaded resources;
• top assumptions, risks, and required decisions.

Each indicator should have an owner, data source, refresh frequency, definition, threshold, and prescribed action. Traffic-light colors should not conceal uncertainty. A green indicator based on incomplete time sheets is not reliable evidence of control.

A strong dashboard distinguishes:

1. Leading indicators: backlog age, unresolved dependencies, capacity gaps, defect arrival, and decision latency.
2. Lagging indicators: actual cost, completed milestones, CPI, and final throughput.
3. Diagnostic indicators: rework hours, waiting time, overtime, and skill-specific utilization.

Dashboards should allow drill-down from portfolio to project, work package, activity, and resource category. However, aggregation can hide local bottlenecks; therefore, summary indicators should be accompanied by exception reports.

9. Reforecasting a Resource-Feasible Baseline

A reforecast is not an informal replacement of the baseline. The following disciplined process is recommended:

1. Freeze the status date. Prevent late entries from changing the reporting boundary without traceability.
2. Validate actuals. Reconcile timesheets, invoices, accruals, progress evidence, and approved changes.
3. Assess physical progress. Confirm that EV is supported by completed and accepted outputs.
4. Re-estimate remaining work. Use task-level estimates, current productivity, known defects, dependencies, and risk.
5. Load the remaining work. Apply the required skill, calendar, location, and availability constraints.
6. Level the schedule. Resolve overloads through sequencing, additional capacity, scope decisions, or milestone negotiation.
7. Calculate cost and time forecasts. Compare bottom-up results with independent EVM and ES estimates.
8. Document assumptions. Record productivity rates, staffing dates, overtime, learning effects, and uncertainty.
9. Approve changes through governance. Distinguish a forecast from a formally rebaselined plan.
10. Monitor corrective actions. Assign owners and dates; do not treat an action as complete merely because it was proposed.

A resource-feasible forecast must satisfy both temporal and capacity constraints. A date is not credible if it requires unavailable specialists, unauthorized overtime, or simultaneous assignment of the same person to conflicting critical tasks.

10. Applied Case: The Northstar Clinical Data Platform

The following is an illustrative case.

Background

Northstar is developing a clinical data platform over a planned 24-week period. The project has a BAC of $1.2 million and requires software engineers, data architects, validation specialists, and clinical reviewers. At week 12, management reports that the project is approximately 70% complete.

Problem

The formal measures show PV = $600,000, EV = $540,000, and AC = $675,000. Two data architects are shared with another project, and validation work has accumulated in a queue. The project manager must determine whether the completion forecast is credible and what control action is required.

Analysis

The indices are:

Cost variance is −$135,000 and schedule variance is −$60,000. A continuation forecast based on CPI is:

This implies a potential overrun of $300,000. However, the project team discovers that $90,000 of AC relates to a one-time data migration purchase and that several hours were incorrectly charged to the project. The CPI therefore contains an atypical component.

A bottom-up review estimates $435,000 of remaining cost. The direct ETC forecast is therefore:

This is substantially lower than the CPI-continuation forecast, but it is credible only if the remaining work can be staffed. Capacity analysis shows that the validation queue requires 480 specialist hours over the next six weeks, while only 300 hours are available. The schedule is therefore not resource-feasible.

Approach and solution

The project manager takes four actions:

1. Corrects the cost coding and separates the one-time migration purchase from recurring productivity trends.
2. Secures 120 hours of temporary validation capacity and reallocates 60 hours from a lower-priority project.
3. Limits new development WIP until validation throughput recovers.
4. Rebuilds the remaining schedule using the actual availability of architects and validation specialists.

The forecast cost remains $1.11 million, but the completion date moves from week 24 to week 26. The sponsor accepts the date change because it is supported by explicit capacity evidence and avoids unapproved overtime.

Outcome and learning points

The case demonstrates that an index alone does not provide a management answer. CPI signaled a serious problem, but data classification showed that part of the variance was non-recurring. The schedule problem was ultimately a bottleneck in validation capacity, not simply a general shortage of labor. The resource-feasible reforecast connected cost, time, throughput, and skill-specific capacity.

11. Limitations, Governance, and Ethics

Metrics can be gamed. Teams may delay recording costs, claim progress before acceptance, split work into artificially small units, or avoid difficult tasks to improve throughput. Such behavior damages decision quality and can create contractual, financial, or regulatory exposure.

Governance should require:

• documented definitions for PV, EV, AC, utilization, and throughput;
• auditable progress evidence;
• segregation of baseline approval and performance reporting where practical;
• controlled access to timesheets and cost data;
• transparent treatment of scope changes and contingency;
• privacy protection for individual performance data;
• review of algorithmic or AI-generated forecasts by accountable managers.

AI and advanced analytics can detect anomalous time charging, forecast demand, identify bottlenecks, and simulate staffing alternatives. Their outputs remain dependent on data quality and model assumptions. An AI forecast should show its input period, confidence or uncertainty range where available, major drivers, and human approval status. It should not be used to infer individual misconduct from utilization data alone.

12. Best Practices

1. Establish a stable and traceable performance measurement baseline.
2. Use objective, acceptance-based earning rules for significant work.
3. Separate effort, cost, value, duration, and capacity rather than treating them as interchangeable.
4. Combine EVM with earned schedule and network-based schedule analysis.
5. Forecast remaining work independently when conditions have changed.
6. Analyze capacity by skill and time period, not only in aggregate.
7. Treat utilization near 100% as a risk signal rather than an automatic success.
8. Use throughput and WIP together; never reward activity without acceptable output.
9. Present leading, lagging, and diagnostic indicators on dashboards.
10. Record assumptions, uncertainty, and corrective-action ownership.
11. Protect the integrity and privacy of resource data.
12. Rebaseline only through formal change control; do not erase unfavorable history.

Chapter Summary

Monitoring and control integrate baseline data, actual effort, remaining work, earned value, schedule position, resource capacity, and forecast information. PV, EV, and AC support cost and schedule variance analysis, while CPI and SPI summarize efficiency. Earned schedule improves time-based interpretation by expressing progress in baseline time units, although it remains an approximation. Utilization, throughput, WIP, and capacity variance reveal bottlenecks and resource-feasibility risks that traditional EVM may miss. EAC and completion-date forecasts must be selected according to their assumptions and reconciled with a bottom-up estimate of remaining work. Effective dashboards connect indicators to thresholds, owners, and decisions. Ethical governance requires auditable progress, transparent assumptions, privacy protection, and human accountability for analytical and AI-supported forecasts.

Key Terms

Actual Cost (AC): Cost incurred for work performed by the status date.

Actual effort: Labor time recorded against project work.

Budget at Completion (BAC): Total approved budget for the performance measurement baseline.

Capacity variance: Difference between planned resource demand and available capacity.

Cost Performance Index (CPI): EV divided by AC; an indicator of cost efficiency.

Earned Schedule (ES): Baseline time at which the current earned value should have been achieved.

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

Estimate at Completion (EAC): Forecast total cost or duration at project completion.

Estimate to Complete (ETC): Forecast cost or effort required to finish remaining work.

Planned Value (PV): Budgeted value of work scheduled by the status date.

Remaining work: Effort required to complete defined outstanding scope.

Schedule Performance Index (SPI): EV divided by PV.

Throughput: Rate of completing acceptable units of work.

Utilization: Productive or assigned effort divided by defined available capacity.

Variance at Completion (VAC): BAC minus EAC for a cost forecast.

Work in Progress (WIP): Work started but not yet completed and accepted.

Resource-feasible baseline: A baseline that can be delivered using available resources and stated constraints.

Reforecast: Updated prediction of cost, duration, effort, and resource demand based on current information.

Review Questions

Short Answer Questions

1. Distinguish planned value, earned value, and actual cost.
2. Why is schedule variance in EVM not the same as schedule delay in weeks?
3. Define remaining work and explain why it may differ from the original estimate minus actual effort.
4. What does a utilization rate measure, and why must its denominator be defined?
5. State two limitations of using CPI as a completion-cost forecast.

Long Answer Questions

1. Explain the complete control cycle from baseline measurement through corrective action and reforecasting.
2. Compare EVM and earned schedule, including their assumptions and limitations.
3. Discuss how throughput, WIP, utilization, and capacity constraints interact in resource control.
4. Design a project dashboard for a resource-intensive project and justify each indicator.
5. Explain the governance and ethical controls required for reliable resource analytics.

Analytical / Application Questions

1. A project has BAC = $500,000, PV = $300,000, EV = $270,000, and AC = $360,000. Calculate CV, SV, CPI, SPI, and the CPI-continuation EAC. Interpret the results.
2. At week 8, EV = $160,000, PV at week 7 = $140,000, and PV at week 8 = $180,000. Calculate ES and SPI_t. Explain what additional evidence is needed before changing the completion date.
3. A team has 1,200 net capacity hours in a month and records 1,080 productive hours. Its remaining backlog requires 1,350 hours, including 300 hours of scarce specialist work for which only 180 hours are available. Diagnose the capacity issue and recommend a resource-feasible control response.

Multiple Choice Questions with Answers

Question 1

Which measure represents the budgeted value of work actually completed?

A. Actual Cost (AC)

B. Planned Value (PV)

C. Earned Value (EV)

D. Budget at Completion (BAC)

Correct Answer: C. Earned Value (EV)

Question 2

A project has EV = $80,000 and AC = $100,000. What is its CPI?

A. 0.80

B. 1.20

C. $20,000

D. $180,000

Correct Answer: A. 0.80

Question 3

Which statement about EVM schedule variance is correct?

A. It is always measured in calendar weeks.

B. It is expressed as EV minus PV and uses value units.

C. It equals actual duration minus planned duration.

D. It remains negative after every late project is completed.

Correct Answer: B. It is expressed as EV minus PV and uses value units.

Question 4

If BAC = $300,000 and CPI = 0.75, what is the CPI-continuation EAC?

A. $225,000

B. $300,000

C. $375,000

D. $400,000

Correct Answer: D. $400,000

Question 5

Which formula defines earned schedule using linear interpolation?

A. ES = EV ÷ AC

B. ES = C + (EV − PV_C) ÷ (PV_{C+1} − PV_C)

C. ES = BAC ÷ CPI

D. ES = AT ÷ PD

Correct Answer: B. ES = C + (EV − PV_C) ÷ (PV_{C+1} − PV_C)

Question 6

A team records 720 productive hours against 800 net available hours. What is utilization?

A. 80%

B. 90%

C. 110%

D. 1,520 hours

Correct Answer: B. 90%

Question 7

What does a positive capacity variance usually indicate when calculated as planned demand minus available capacity?

A. Excess available capacity

B. A resource shortage

C. A favorable cost variance

D. Completed work exceeding planned work

Correct Answer: B. A resource shortage

Question 8

Under Little’s Law, if throughput remains stable and WIP increases, what generally happens to lead time?

A. It decreases to zero.

B. It increases.

C. It becomes unrelated to WIP.

D. It equals CPI.

Correct Answer: B. It increases.

Question 9

Which is the most appropriate reason to prefer a bottom-up ETC over a simple CPI-based EAC?

A. The project has no remaining work.

B. The original cost efficiency is known to apply unchanged forever.

C. Scope, resource mix, or operating conditions have materially changed.

D. PV is equal to EV.

Correct Answer: C. Scope, resource mix, or operating conditions have materially changed.

Question 10

What is the main purpose of a resource-feasible reforecast?

A. To erase unfavorable historical performance.

B. To produce a forecast that respects available skills, time, dependencies, and constraints.

C. To maximize utilization to exactly 100%.

D. To replace all approved baselines without change control.

Correct Answer: B. To produce a forecast that respects available skills, time, dependencies, and constraints.