Schedule Analysis and Performance
How to Perform Schedule Variance Analysis
Compare the approved plan with actual progress and the current forecast, identify why dates differ, determine which differences matter, and select a response supported by schedule logic rather than surface-level numbers.
Meaning
A variance is a signal to investigate—not a conclusion
Schedule variance analysis examines differences between a selected reference plan and actual or forecast schedule results. A milestone may finish later than its baseline date, an activity may start earlier than planned, a task may take longer than its baseline duration, or a current completion forecast may move from one reporting period to the next. Each difference answers a different question.
A positive or negative number alone does not explain project health. A late activity may have enough float to avoid affecting any controlled milestone. An early activity may create no benefit because its successor cannot start. A finish milestone may remain unchanged because available float absorbed a delay, while a formerly comfortable path becomes near-critical. Conversely, a milestone variance may result from corrected schedule logic rather than recent execution failure.
Schedule variance analysis is also distinct from earned value schedule variance. Date-based variance compares planned, actual, and forecast dates or durations. Earned value schedule variance compares budgeted value of work performed with budgeted value of work scheduled. Both can support management, but their units, meanings, and limitations differ. This guide focuses on date- and network-based schedule variance.
Reference dates
Choose the correct comparison before calculating anything
| Comparison | Question answered | Typical use |
|---|---|---|
| Actual versus baseline | Did completed work occur when the approved plan expected? | Historical execution performance |
| Current forecast versus baseline | Where is incomplete work now expected relative to the approved plan? | Commitment and downstream-impact analysis |
| Current forecast versus previous forecast | Has the outlook improved or deteriorated since the last update? | Period-to-period trend analysis |
| Current forecast versus target | Can the latest forecast meet a management objective? | Recovery and scenario planning |
| Actual duration versus baseline duration | Did completed work take longer or less time than planned? | Estimating feedback and productivity review |
| Remaining duration versus prior remaining duration | Has the estimate to complete changed? | Forecast-quality assessment |
State the reference explicitly in every report. “Milestone is ten days late” is ambiguous unless the reader knows whether the comparison is against the original baseline, an approved revised baseline, last month’s forecast, a contractual commitment, or an internal target. Several comparisons may be useful, but they should not be blended into one number.
Confirm that the baseline is the approved controlled version and that the current schedule has a valid status date. Confirm that the calendars used to calculate variance are known. Ten calendar days, ten working days on a five-day calendar, and ten shifts on a continuous calendar are not interchangeable.
Core calculations
Calculate schedule variances consistently
Start-date variance
Actual or forecast start − baseline startA positive result under the convention used here means later than baseline; a negative result means earlier. State the convention because some systems reverse the sign.
Finish-date variance
Actual or forecast finish − baseline finishUse this for activities and milestones, but interpret the result through successors, path position, float, and the milestone being controlled.
Duration variance
Actual or current duration − baseline durationCheck that durations use comparable calendars and definitions. Elapsed time and working duration can produce different results.
Forecast movement
Current forecast − previous forecastThis measures recent movement, not total baseline variance. A milestone can improve this month while remaining substantially later than baseline.
Schedule growth
(Current duration − baseline duration) ÷ baseline duration × 100This provides scale, but should not replace the underlying date and path analysis. A percentage can hide where growth occurred.
Float movement
Current total float − prior or baseline total floatFalling float can provide early warning before a milestone moves. Compare float only when calculation settings, calendars, logic, and status conditions are compatible.
Analysis workflow
A twelve-step method for credible variance analysis
Define the decision the analysis must support
Decide whether the purpose is milestone assurance, monthly reporting, recovery planning, supplier review, resource prioritization, or lessons learned. The purpose determines the appropriate level of schedule detail and comparison periods.
Freeze and identify the comparison versions
Record the baseline identifier, current schedule revision, previous update revision, status date, extraction time, and applicable calendars. Preserve the analyzed versions so the result can be reproduced later.
Validate schedule quality before trusting the variance
Review missing logic, invalid actual dates, open-ended activities, inappropriate constraints, unusually long durations, stale status, and broken calendars. Variance calculated from an unreliable schedule may be mathematically correct but managerially misleading.
Build a complete variance population
Include completed, in-progress, and not-started work. Identify activities that started or finished early, on time, or late; work that should have started or finished but did not; and work that occurred earlier than planned. Avoid reviewing only visibly late milestones.
Apply materiality thresholds
Establish thresholds suited to the project’s scale and commitments. Consider date movement, percentage growth, float consumption, path position, external effect, and repeated small variance. A small deviation on a critical handoff may matter more than a large deviation on non-driving work.
Trace variance through schedule logic
Follow predecessors to identify where movement originated and successors to identify what it affects. Determine whether float absorbed the movement, another path became controlling, or a constraint prevented natural propagation.
Separate execution variance from model variance
Execution variance arises from events such as late work, changed productivity, or delayed approvals. Model variance may arise from corrected logic, changed calendars, revised durations, new scope, or altered constraints. Both matter, but they require different responses and explanations.
Identify direct cause and root cause
“Testing started late” is a direct schedule observation. The root cause may be incomplete requirements, unavailable environments, failed entry criteria, resource contention, or late supplier delivery. Continue investigating until the explanation supports a practical management response.
Test downstream consequence
Identify affected handoffs, milestones, resource windows, procurement dates, reviews, and external commitments. Determine whether the variance changes only local timing or creates a credible threat to a controlled outcome.
Review the trend across reporting periods
One-period movement can reflect timing noise or a corrected forecast. Repeated movement in the same direction is more informative. Track milestone dates, float, incomplete work, and unresolved causes across updates to identify persistent deterioration or sustained recovery.
Select a proportionate response
Responses may include monitoring, correcting data, obtaining missing decisions, reallocating resources, resequencing work, managing a dependency, analyzing a recovery scenario, or initiating change control. Do not prescribe acceleration before confirming feasibility and side effects.
Record ownership and verify the next update
Assign an accountable owner, action, target date, expected schedule mechanism, and review point. At the next status cycle, test whether the action occurred and whether the expected path or milestone effect materialized.
Interactive calculator
Milestone variance and trend analyzer
Enter dates for one milestone. The calculator reports baseline variance, recent forecast movement, direction, and a structured interpretation prompt. It uses calendar-day differences; use your scheduling system when working-day calculations are required.
Cause analysis
Move beyond “the task took longer than planned”
A useful variance explanation connects the schedule symptom to an actionable cause without overstating certainty. Start with verified chronology, examine the planning assumptions, and determine whether the schedule accurately represented the work.
Execution causes
- Lower productivity than assumed
- Late start or interrupted work
- Quality failure and rework
- Resource unavailability
- Delayed decision or approval
- Supplier or interface failure
Planning causes
- Duration estimate was unsupported
- Scope was incomplete
- Dependency was omitted
- Resource limits were ignored
- Calendar assumption was unrealistic
- Acceptance work was not modeled
Schedule-model causes
- Missing or incorrect logic
- Inappropriate constraint
- Changed lag or relationship
- Incorrect progress status
- Calendar or software-setting change
- Baseline mapping problem
More than one cause may contribute. Avoid assigning an exact portion of delay to each cause unless the evidence and analysis support that allocation. Where responsibility is disputed, report the confirmed schedule condition, document competing explanations, and identify the unresolved decision or investigation.
Interpretation
Four common variance patterns and what they mean
Late against baseline, improving against prior forecast
The milestone remains behind its approved plan but has recently recovered time. Report both facts. Declaring it “back on track” may be misleading unless the current forecast meets the controlling commitment.
Finish date stable, float declining
The commitment has not moved, but protective flexibility is being consumed. Investigate the path and remaining uncertainty before the milestone begins to slip.
Activity late, milestone unaffected
Available float or an independent sequence may absorb the variance. Confirm that the activity is not driving another handoff, resource window, or lower-level commitment.
Milestone moved after schedule correction
The new forecast may be more credible even though it is later. Explain the defective logic or assumption, the correction, and the resulting impact separately from current execution performance.
Early work is not automatically favorable
Work performed earlier than baseline can indicate good execution, but it can also reflect unauthorized resequencing, premature work, resource diversion, or missing predecessor logic. Confirm that required inputs were available and that early completion benefits downstream work rather than creating inventory, rework, or acceptance problems.
Worked example
Analyze the path, not only the late date
Assume a design-approval milestone had a baseline date of April 8, a previous forecast of April 12, and a current forecast of April 19. Its baseline variance is eleven calendar days late, while its most recent forecast movement is seven days later.
The approval drives procurement release. Procurement has a fixed supplier production window, so a late release moves equipment delivery. Equipment delivery drives installation, but the installation milestone originally had six days of total float. The eleven-day approval variance therefore does not translate automatically into eleven days of project completion delay. Part may be absorbed by available float, and the remainder depends on supplier and installation logic.
Investigation shows that the approval package was submitted on time but returned because a required interface analysis was incomplete. The direct cause is failed approval. The deeper planning cause is that the interface analysis was not represented as a predecessor activity. Corrective action should therefore address both the current approval package and the missing planning logic. Simply shortening installation would not correct the originating weakness.
The final analysis should state the baseline variance, recent movement, cause evidence, affected path, available float, forecast consequence, corrective owner, and review date. It should distinguish confirmed effects from recovery assumptions.
Response planning
Select corrective action according to the diagnosed cause
| Finding | Possible response | Required caution |
|---|---|---|
| Incorrect or stale status | Validate and correct schedule data | Do not treat a data correction as execution recovery. |
| Missing dependency | Add valid logic and reassess the path | Follow applicable schedule change controls. |
| Late external decision | Escalate the decision with a required-by date | Identify the authorized decision maker and consequence. |
| Resource contention | Reprioritize, reassign, or revise sequence | Test the effect on work losing those resources. |
| Unrealistic duration estimate | Re-estimate remaining work using evidence | Preserve baseline variance and document the basis. |
| Recoverable near-critical movement | Monitor triggers and prepare a response | Avoid unnecessary acceleration that creates new risk. |
| Confirmed critical-path delay | Analyze feasible recovery scenarios | Do not promise recovery until resources and assumptions are approved. |
A response is complete only when it has an owner, target date, expected schedule effect, dependencies, approval state, and next review point. At the next update, compare the actual result with the expected effect. If the variance continues, revisit the diagnosis rather than repeatedly extending the same ineffective action.
Failure prevention
Common schedule variance analysis mistakes
| Mistake | Why it fails | Better method |
|---|---|---|
| Reporting only the final milestone | Early path deterioration remains hidden. | Analyze intermediate milestones, handoffs, paths, and float. |
| Calling every late activity critical | Lateness and criticality describe different conditions. | Verify path position and downstream effect. |
| Using an uncontrolled baseline | The reference may have changed, invalidating comparisons. | Use the approved version and preserve identifiers. |
| Ignoring calendar differences | Variance magnitudes become inconsistent. | State whether results use working or calendar days. |
| Explaining a symptom as a cause | “Started late” does not explain why it started late. | Investigate decisions, inputs, resources, assumptions, and logic. |
| Hiding favorable variance | Early work may expose planning or sequencing problems. | Review favorable and unfavorable exceptions. |
| Looking at one reporting period | A persistent trend may be missed. | Maintain comparable measures across update cycles. |
| Claiming recovery from a scenario | A modeled option may lack approval or resources. | Separate current forecast, scenario, and approved response. |
Review checklist
Validate the analysis before publishing it
- The baseline, current schedule, previous forecast, status date, and calendars are identified.
- The analyzed schedule passed relevant logic and status-quality checks.
- Calendar-day and working-day measurements are clearly distinguished.
- Start, finish, duration, milestone, forecast, and float differences are not mixed together.
- Materiality reflects path position, commitments, and downstream effect—not magnitude alone.
- Each important variance was traced through predecessors and successors.
- Execution effects and schedule-model changes are explained separately.
- Direct cause, contributing factors, and root cause are not confused.
- Trend direction across multiple updates has been reviewed.
- Facts, assumptions, scenarios, and unresolved explanations are labeled.
- Corrective actions have owners, dates, dependencies, and expected effects.
- The next update includes a test of whether the response worked.
Frequently asked questions
Schedule variance analysis FAQs
Does schedule variance always mean the project will finish late?
No. Variance means performance or forecast timing differs from the selected reference. Available float, parallel paths, successor logic, corrective action, and other factors determine whether the project completion milestone is affected.
Should variance be calculated in calendar days or working days?
Use the convention defined for the schedule and report it clearly. Working-day variance is often useful for execution analysis, while calendar days may help external audiences. Do not compare numbers produced under incompatible calendars without adjustment.
What variance threshold should trigger investigation?
No single threshold fits every project. Combine date movement with path position, float, milestone importance, external commitment, recurrence, risk, and project scale. Define thresholds in the schedule-management process before exceptions occur.
Can an activity be late without having negative float?
Yes. It may be later than baseline but still retain positive float. Baseline variance measures deviation from plan, while total float describes flexibility relative to a schedule endpoint under current logic.
Why compare current forecast with both baseline and previous forecast?
Baseline comparison shows total deviation from the approved plan. Previous-forecast comparison shows recent movement. Together they reveal whether a late milestone is stabilizing, improving, or continuing to deteriorate.
What if the baseline schedule was poor quality?
Disclose the limitation. Correcting or replacing an unreliable baseline may require controlled approval, but the project should not present questionable variance as precise performance evidence. Preserve historical records and explain the treatment used.
Is favorable variance always good?
No. Early work may be genuinely beneficial, but it can also indicate missing logic, unauthorized resequencing, premature execution, resource diversion, or work completed before necessary inputs were mature. Investigate material favorable exceptions.
How often should variance analysis be performed?
Perform it after each controlled status update at a frequency suited to the work and decision cycle. High-risk or fast-moving activities may need closer monitoring than the formal reporting frequency.
Create and revise practical plans with the verified Weekly Schedule Builder. More focused scheduling methods are available in ScheduleFlowLab Guides.