Skip to content

Plan Project Delivery and Integrate the Management Plan

Step 18 of 57

PMP Integrated Project Management Plan: Estimates, Baselines, and Decisions

Introduction

An integrated project management plan brings scope, schedule, resources, and cost into agreement. If a supplier cannot deliver until June, the testing schedule and staffing plan need to reflect that dependency.

This lesson continues Task 1 in the Process domain of the July 2026 PMP Examination Content Outline (ECO). Process accounts for 41% of exam coverage. After choosing a project development approach, you will estimate effort and capacity, reconcile plans and baselines, and use verified data to support decisions.

Learning objectives

  • Explain what the integrated project management plan contains and how subsidiary plans relate to it.
  • Apply estimating techniques (analogous, parametric, bottom-up, three-point, relative sizing) to effort and resources.
  • Review a consolidated plan for dependencies, gaps and business value.
  • Maintain the plan through integrated change control and regular reviews.
  • Use data to decide, and evaluate the output of AI-assisted analysis critically.

1. The integrated project management plan

The project management plan is the single approved document (or set of linked documents and tools) describing how the project will be executed, monitored and controlled. Integration means checking how a decision affects the other parts. When scope changes, review schedule, cost, resources, quality, risk, and procurement, then update the parts affected by the approved change.

Component Content
Subsidiary plans Scope, requirements, schedule, cost, quality, resource, communications, risk, procurement, stakeholder engagement (and, in adaptive work, release and iteration plans)
Baselines Scope baseline (scope statement, WBS, WBS dictionary), schedule baseline, cost baseline
Other components Change management plan, configuration management plan, performance measurement baseline, project life cycle and development approach, management reviews

Project documents such as the risk register, issue log, and requirements records support the plan. Update them as information changes, following the project’s document controls. An updated register does not itself authorise a baseline change.

An adaptive project may use a roadmap, release forecasts, iteration plans, and a backlog. The amount of detail depends on the project’s risks and commitments. In a hybrid project, the predictive frame (milestones, budget, governance) is integrated with the adaptive plans (backlog, iteration cadence). Integration is the project manager's role; no subsidiary plan owner can do it alone.

Example: a plan that does not integrate

A team builds a hospital scheduling system. The schedule assumes the data-migration specialists start in March. The resource plan shows them on another project until June. The cost plan funds only 2 months of their work, and the risk register does not mention the conflict. Each plan has a different assumption about the same work. Integration review would have revealed it before execution.

2. Estimating work effort and resource requirements

Effort is the amount of work, for example in person-hours. Duration is the working time over which it takes place, using the relevant calendar. Labour cost depends on effort and rates. The full cost may also include equipment, materials, and other expenses.

Technique How it works Use when Limit
Analogous Uses the actuals of a similar past project, adjusted Early, little detail Reliability depends on the similarity and quality of the historical data
Parametric Applies a unit rate to a measure (cost per square metre, hours per screen) Data from history, scalable work Requires a supported relationship between the unit and effort or cost
Bottom-up Estimates each work package, then sums Detailed scope known Time-consuming
Three-point Combines optimistic (O), most likely (M) and pessimistic (P) values Uncertainty Needs good inputs
Expert judgement / Delphi Experts estimate; anonymous rounds converge Novel work Bias
Relative sizing Story points, T-shirt sizes, affinity grouping, planning poker Agile backlogs Use a team-specific forecast with observed delivery data if a date estimate is needed
Reserve analysis Adds contingency for identified risks Any Must be justified

Three-point formulas. The triangular mean is (O + M + P) / 3. The conventional PERT weighted estimate is (O + 4M + P) / 6. A traditional PERT approximation for the standard deviation is (P − O) / 6.

Worked example (PERT). A task has O = 10, M = 14, P = 24 days. PERT estimate = (10 + 56 + 24) / 6 = 15 days; standard deviation = (24 − 10) / 6 = 2.33 days. Triangular = 16 days. Use 2.33 days as a rough measure of spread under that model. To make a probability statement about finishing by a date, use a justified distribution and its assumptions.

Progressive elaboration means refining estimates as you learn more. State the range and assumptions that the evidence supports. The range may narrow as uncertainties are resolved or widen when a new risk is discovered.

Estimating in agile projects

Some teams use story points to estimate relative size and track velocity, the points completed in an iteration. Scrum does not prescribe story points or velocity. If the remaining backlog is 240 points and the average velocity is 30 points per two-week sprint, the forecast is 240 / 30 = 8 sprints, 16 weeks. Velocity belongs to the team and must not be used to compare teams or to pressure them. That forecast assumes the backlog and capacity stay comparable. A range based on recent comparable Sprints can show sensitivity to different rates. Label it as a scenario range and revise it when the scope or team changes.

Estimating resources

Resource requirements include people (skills, quantity, time), equipment, materials, licences, facilities and tools. Consider productivity, learning curves, availability, calendars, and time for onboarding, mentoring, and coordination. Some work cannot be done in parallel.

Worked example: from effort to resources

Statement. A backlog for a client portal is estimated at 1,800 person-hours of build and test work. Each of four developers has 5.5 hours per working day available for build and test after allowing for meetings and support. The sponsor asks when it will be ready and how many people are needed to finish in 12 weeks (60 working days).

Reasoning.

  1. Capacity of 4 developers: 4 × 5.5 × 60 = 1,320 hours in 12 weeks.
  2. Gap: 1,800 − 1,320 = 480 hours.
  3. Options: add people, extend the time, or reduce scope. To finish in 12 weeks with the same productivity: 1,800 / (5.5 × 60) = 5.45, so six developers at that availability, before allowing for onboarding or work that cannot run in parallel.
  4. Allow for onboarding: assume the two new developers contribute no build or test work during their first ten days. Their remaining capacity is 2 × 5.5 × 50 = 550 hours. If mentoring takes 80 hours from the existing team, total capacity is 1,320 + 550 − 80 = 1,790 hours.
  5. Recommendation: present the three options with cost and risk, and let the sponsor decide.

Why not simply add two developers? Under the stated assumptions, the extra people leave a gap of ten hours before other risks. Show the effect of different onboarding assumptions and confirm that the extra work can be parallelised.

3. Assessing consolidated plans: dependencies, gaps and business value

After assembling the plan, challenge it as a whole.

Dependencies. Four types between activities: finish-to-start, start-to-start, finish-to-finish, start-to-finish. These relationships describe the order of starts and finishes. A dependency can also be classified by whether it is mandatory or discretionary and whether it is internal or external.

  • Mandatory (hard logic): physical or contractual necessity.
  • Discretionary (soft logic): best practice or preference.
  • External: outside the project (regulator approval, supplier delivery).
  • Internal: within the team's control.

Also look for dependencies between projects, with operations (a change freeze at year end), shared resources, and between plans (a procurement lead time that affects the schedule).

Gaps. Missing items, such as requirements without a work package, risks with no response in the schedule, training with no budget, acceptance criteria with no tester, a stakeholder group with no communication.

Business value. The plan must still deliver the benefits that justify the project. Check periodically: is the business case still valid, are the benefits still achievable, has the market, regulation or priority changed? If the value has eroded, the right action is to raise it to the sponsor and governance body, who can continue, change or stop the project.

A practical review checklist:

Question Where to look
Can each applicable requirement be traced to the work and a way to verify it? Requirements traceability records, work breakdown structure or backlog, and test plan
Do dates, resources and costs agree? Schedule, resource and cost plans
Are risks reflected in reserves and activities? Risk register, baselines
Are external dependencies and lead times in the schedule? Procurement plan
Are quality and acceptance activities included? Quality plan
Does the plan still meet the business case? Business case, benefits plan

Example: a dependency gap found at consolidation

For a retail opening, the authority needs 15 working days from a valid request to inspection, and stock delivery must wait for the resulting certificate. The schedule has only ten working days between the earliest valid request and planned stock delivery. Add the missing request and the inspection and certificate dependencies to the forecast. Check the authority’s calendar and processing time, then present options to the sponsor. Change the baseline only with approval.

4. Maintaining the plan

A plan is a living tool. Maintain it by:

  • Comparing actuals to the baseline and analysing variances at each reporting cycle.
  • Using integrated change control: changes to baselines are submitted, assessed for all impacts, approved or rejected by the right authority, then implemented and communicated.
  • Updating the plan and documents after approved changes, after risk events, at stage gates and after retrospectives.
  • Keeping version control and a record of decisions.
  • Re-baselining only through approval, when the old baseline is no longer a meaningful measure (for example after a large approved scope change); do not re-baseline to hide a poor performance.
  • Reviewing the approach and tailoring when results show that something is not working.

In adaptive projects, the backlog is refined continuously, release plans are updated after reviews, and the roadmap changes with feedback; use the authority agreed for each decision. Reordering work may be within the Product Owner’s remit in Scrum. Changes to baselines or contractual commitments use their applicable approval processes. See defining and managing project scope for scope decisions and traceability.

Corrective action addresses a performance deviation. Preventive action reduces the chance of a future deviation. Use a formal change request for changes to controlled deliverables, documents, or baselines when required by the project’s process.

5. Data-driven decisions and AI-assisted analysis

The ECO states that the project manager makes data-driven decisions. This means decisions rest on measured information — status data, forecasts, trends, risk analyses, benefits tracking — and not on assumptions or the loudest opinion.

Good practice:

  1. Define the decision first, then the data that would change it.
  2. Check data quality: source, completeness, timeliness, definitions, bias.
  3. Use suitable analysis: trend, variance, earned value, Monte Carlo simulation, sensitivity analysis, cumulative flow, velocity ranges.
  4. Present options with consequences, not only numbers.
  5. Record the decision and its rationale, and revisit if the data changes.

AI in project management

The July 2026 outline lists AI as part of the context in which projects run. AI tools can summarise status, draft reports and plans, forecast completion dates, detect schedule or budget anomalies, classify risks, and analyse lessons learned. The project manager remains accountable for the result.

When evaluating an AI-supported analysis:

  • Check the data used (is it complete and current? is confidential data allowed to be shared with the tool?).
  • Validate the output against known facts and with a human expert; AI output can be wrong or invented.
  • Understand limits and bias: a forecast trained on past projects may not suit a project with a new approach.
  • Respect policies: privacy, intellectual property, security, organisational rules for approved tools.
  • Keep human judgement for decisions, especially with ethical, legal or people consequences.
  • Document how the tool was used and how its output was verified.

Before acting on an AI forecast, reproduce the calculation with approved data and verify any factual claims used in the decision.

Example: an AI forecast that looks too good

An AI assistant forecasts nine Sprints from an old backlog of 378 points and a rate of 42 points per Sprint. The project manager checks the current backlog of 435 points and the last three completed Sprints at 31, 29, and 33 points after two developers left. At 29–33 points per Sprint, the forecast is about 14–15 whole Sprints if capacity and scope stay comparable. She presents the revised forecast, explains its assumptions, and compares changing scope, staffing, or date with the sponsor.

Worked example: using data to choose between options

Statement. A project is 40% through its schedule. Planned value (PV) is EUR 400,000, earned value (EV) is EUR 340,000 and actual cost (AC) is EUR 380,000. The sponsor asks you whether to ask for more budget.

Reasoning.

  1. Cost performance index CPI = EV / AC = 340 / 380 ≈ 0.895. Schedule performance index SPI = EV / PV = 340 / 400 = 0.85. Keep the unrounded ratio for subsequent calculations.
  2. The work performed has cost more than its budgeted value, and less budgeted work is complete than planned. If the same cost efficiency continues for the remaining work and the budget at completion (BAC) is EUR 1,000,000, EAC = BAC / CPI = 1,000,000 × 380 / 340 ≈ EUR 1,117,647. That is about EUR 118,000 over BAC. The SPI alone gives no completion date.
  3. Before asking for funds, find the cause (rework, staffing, wrong estimates, scope growth) and options (re-plan, reduce scope, add capacity, improve quality), then present the forecast and options with their consequences to the sponsor, using the change process if the baseline is to change.

Why not ask for money straight away, or ignore it? A request without analysis lacks credibility; ignoring the variances breaks governance and transparency.

6. Common pitfalls and exam tips

Pitfalls

  • Treating subsidiary plans as separate documents owned by different people with no integration.
  • Estimating with a single number, no range, no reserves, no stated assumptions.
  • Using one technique for everything, or relying on the estimate of the most optimistic person.
  • Adding people to a late project without allowing for onboarding.
  • Never revisiting the business case after approval.
  • Re-baselining to hide poor performance.
  • Updating the plan informally, without change control.
  • Trusting AI output without verification, or sharing confidential data with an unapproved tool.

Exam tips

  • "Integrated" means checking the impact of a change on all plans and baselines; the best answer often begins with analysing the impact.
  • Choose the estimating technique from the context: early and little detail, analogous; historical data, parametric; detailed scope, bottom-up; uncertainty, three-point; agile backlog, relative sizing.
  • Use the PERT weighted estimate when the question specifies it. Use EAC = BAC / CPI when the remaining work is expected to continue at the observed cost efficiency.
  • Changes to baselines go through integrated change control, never by informal agreement.
  • Business value is checked throughout; if it has disappeared, escalate to the sponsor.
  • With AI, the best answer validates the output, protects data and keeps human accountability.
  • Prefer facts and data over opinions and assumptions.

7. Mini case study

Situation. You manage the replacement of the booking platform of a ferry operator, 14 months, EUR 5.2 million, hybrid approach: milestones for regulatory approval and port-system integration, two-week sprints for the customer web app. After month 5 you consolidate the plans for the first major review. Findings: the port-system integration depends on a supplier release promised for month 8, but the schedule shows integration testing in month 7; the budget does not include a security penetration test that the new data-protection policy requires; two senior developers are planned for 100% on this project but also on a maintenance contract at 40%; the backlog has 520 points, velocity is 26 points per sprint, with 22 sprints left in the schedule; a new competitor has launched a similar app, and the marketing director now asks for a loyalty module.

Question. What do you do?

Model solution.

  1. Calculate the forecast. At the previous delivery rate of 26 points per Sprint, 520 points would take 20 Sprints. Adding 60 points would take about 22.3 Sprints, or 23 whole Sprints against 22 available. Treat both forecasts as provisional until staffing and dependencies are reconciled.
  2. Resolve dependencies. Move integration testing after the supplier release or negotiate an earlier test version; add the dependency to the schedule and assess the risk to the milestone.
  3. Fill gaps. Estimate and schedule the penetration test and its acceptance criteria in the proposed plan. Seek approval for the cost baseline change.
  4. Correct resources. Reflect the two developers’ 60% availability and revise the forecast with the team. Compare backfill, a later date, and reduced scope.
  5. Re-check value. Meet the sponsor and product owner: check whether the competitor changes the assumptions behind the business case. Evaluate the loyalty module against the other backlog items by value and cost, not as an automatic addition.
  6. Use change control. Prepare change requests for baseline changes (schedule, cost), with analysis of impacts on all plans and risk; get decisions from the steering committee.
  7. Communicate and update. Update the integrated plan, the risk register, the stakeholder communication and the roadmap; record decisions.
  8. Use data and AI carefully. If an AI tool is used to simulate forecasts, validate its input data and assumptions and document them.

Key terms

  • Project management plan: the approved description of how the project is executed, monitored and controlled.
  • Subsidiary plan: a component plan (scope, schedule, cost, quality, resources, communication, risk, procurement, stakeholders).
  • Baseline: the approved version of scope, schedule or cost against which performance is measured.
  • Analogous / parametric / bottom-up estimating: techniques based on similar projects, unit rates, or summing detail.
  • PERT: (O + 4M + P) / 6.
  • Velocity: story points completed per iteration.
  • Progressive elaboration: refining estimates and plans as information increases.
  • Integrated change control: assessing and deciding changes across all components of the plan.
  • Data-driven decision: a decision grounded in validated project data.

Summary checklist

  • I assemble an integrated plan whose subsidiary plans and baselines agree with each other.
  • I choose the right estimating technique and give ranges with stated assumptions.
  • I translate effort into resource needs, allowing for productivity and onboarding.
  • I review consolidated plans for dependencies, gaps and continued business value.
  • I maintain the plan through integrated change control and regular reviews.
  • I base decisions on validated data and evaluate AI-assisted analysis critically, protecting data and keeping accountability.

Official sources