Choose critical path software that shows how it identifies critical activities, recalculates when schedule inputs change, and lets a scheduler inspect logic and float. A Gantt chart that colors a chain of bars is not, by itself, evidence that the forecast finish is being calculated from a dependable network. If a fit-out activity slips, you need to see whether its successors and the project finish respond under the schedule’s calendars, constraints, and calculation settings.
For contractual review or delay analysis, shortlist tools with inspectable calculation settings and evidence about float and path behavior. If the main problem is that teams do not keep dates current, evaluate a work-management Gantt tool too, but test its scheduling limits against a representative copy of your schedule. This guide compares documented behavior for five products, not their prices, ratings, or complete feature sets.
The central buying question is simple: can the product produce an explainable schedule result, or does it mainly provide a useful visual highlight? The answer depends on the product, edition, plan, schedule data, and settings you actually use.
What critical path software should calculate
The critical path is the sequence of driving activities that determines a forecast project finish under a specified schedule calculation. A delay to an activity on that path can affect the finish, but the label “critical” is conditional on schedule data, relationship logic, calendars, progress settings, constraints, and the product’s criticality rule. It does not automatically establish contractual criticality.
Total float is the time an activity or sequence can slip before the project finish is affected. Free float is the time it can slip before affecting a connected successor. Path membership says whether an activity belongs to a path identified by a particular method. A tool’s criticality threshold is another separate setting. Record these as distinct facts, rather than combining them into one generic float result.
A colored critical task is a display result. Before relying on it, identify the calculation rule and inspect the underlying logic. If the tool cannot show enough evidence for the decision at hand, treat the highlight as a planning aid, not schedule validation.
For example, Oracle Primavera P6 documents both total-float threshold criticality and a Longest Path option. Those settings can identify different activity sets. P6 also documents multiple float paths as a separate analysis with its own endpoint, path count, and float basis. The distinction matters when a reviewer asks not only which activity is highlighted, but why it was highlighted and what alternative paths were considered.
Compare documented critical-path behavior before comparing brands
The products below expose different kinds of evidence. A longest-path calculation, a float threshold, a dependency-chain highlight, and a network diagram are not equivalent methods. Use the table to form a shortlist, then confirm the actual edition or plan and settings you would use.
| Tool | Documented behavior | Decision to verify |
|---|---|---|
| Primavera P6 | Can identify critical activities by a total-float threshold or Longest Path. It separately supports configurable multiple float paths, including choices such as path endpoint and float basis. | Which criticality rule is selected? Are multiple float paths configured separately? Check constraints and resource leveling when the apparent driving path is unexpected. Oracle P6 critical-activity help and multiple-float-path help. |
| Microsoft Project | Defines critical tasks using slack and allows a configurable threshold. It distinguishes total slack from free slack and can display multiple paths. | Check the threshold, slack fields, and product edition. Project Online is scheduled to retire on September 30, 2026. That notice does not automatically retire Project desktop, Project Server, or Planner. Microsoft critical-path guidance and Project Online lifecycle. |
| ClickUp | Documents Critical Path and Slack Time in Gantt view. Its help warns that time gaps or missing dependency links can shorten the displayed path. | Check whether the Gantt feature fits your review needs and how your team will find and repair missing links. The reviewed documentation does not establish specialist controls such as P6 multiple-float-path analysis. ClickUp Critical Path and Slack Time help. |
| Planera | Documents a longest continuous path, recalculation when the schedule is run, and total-float recalculation. Its help warns that hard constraints can create artificial critical paths. | Test the displayed path against relationships and inspect hard constraints. Planera critical-path calculation help. |
| Wrike | Current help documents Critical Path on Business, Pinnacle, and Apex accounts, with legacy-plan references also shown. The New Gantt Chart can show total and free float. | Confirm plan availability and whether the documented Gantt evidence meets your review requirement. Do not assume the feature is available on Free or Team. Wrike Critical Path help and Wrike plan types. |
These are evidence-based distinctions, not a universal ranking. Vendor help confirms the behavior it documents, not every capability of every edition. Before purchase, verify the current plan, schedule owner workflow, and calculation settings in official documentation or a product trial.
Test whether the calculated path survives a schedule change
A controlled schedule test is more useful than a feature checklist alone. The schedule owner should run the same small test in each shortlisted product, save the input copy and settings, and ask the vendor to explain any surprising result.
Do not expect a fixed one-for-one movement in the finish date. Calendars, constraints, completed work, alternate paths, and calculation settings can change the response. For P6 schedules with out-of-sequence progress, record whether Retained Logic, Progress Override, or Actual Dates was selected. Oracle documents these as distinct scheduling behaviors in its P6 scheduling options.
For each product, the test record should contain the original schedule version, the exact input change, the calculation setting used, the observed finish response, and the reviewer decision. A result that moves in an unexpected direction is not automatically a software defect. It is a request to inspect calendars, constraints, status, relationship logic, and alternate paths.
Check schedule health before believing a result
A calculation can be internally consistent while the schedule is operationally weak. Missing relationships, open ends, unsuitable calendars, hard constraints, or gaps in a driving chain can undermine the interpretation. Review changes between runs as well: a path change may be legitimate, but an unexplained shift deserves investigation.
- Check open ends and missing or broken relationships, allowing for intentional project start and finish activities.
- Identify hard constraints and ask whether each constrained date is approved and supported by network logic.
- Inspect continuity of the proposed driving chain and investigate gaps or time breaks.
- Record total float, free float, and path membership separately where available.
- Compare the critical path and finish forecast with the prior run, then explain material changes.
- Confirm the schedule version, data date, calendar configuration, calculation method, and progress-logic setting.
The U.S. Department of Defense’s Integrated Master Plan and Integrated Master Schedule Guide discusses schedule health and the DCMA 14-point metrics as a framework for identifying problem areas and asking informed questions. The metrics are not a complete validation standard or a replacement for scheduler judgment. For contractual, claims, or recovery-plan use, route material logic gaps, constraint-driven criticality, negative float, or unexplained path changes to a qualified scheduler or project-controls lead.
Also monitor near-critical and low-float branches. Research involving 323 schedules from 77 real-world projects and more than $10 billion in total budget describes delay propagation as a network property rather than an effect confined to one critical path. The practical implication is to watch connected branches that could become driving, not just the currently highlighted chain. See the Scientific Reports study, “Activity networks determine project performance”.
Choose the tool category that matches the schedule’s job
A specialist CPM engine is a stronger candidate when calculation settings, float-path review, and formal project-controls review are central. A work-management Gantt tool may be a better fit when the schedule fails because owners update tasks elsewhere and dates go stale. A network-diagramming tool can help teams communicate dependencies, but a diagram should not be assumed to recalculate execution dates.
Choose the product that preserves the evidence your review audience needs, not merely the product with the most convincing Gantt display.
Use a specialist engine when schedules are contractual deliverables, require detailed calculation settings, or need formal float and path review. Evaluate a work-management tool when schedule upkeep and ownership are the main failure. Evaluate a diagramming tool when the deliverable is network logic rather than an automatically recalculated execution schedule.
Before selecting, answer four operational questions: Is this schedule the system of record? Who updates actuals and remaining duration? How often is it recalculated and reviewed? What calculation evidence must be retained for the audience? Also confirm how baselines are identified and whether the product can preserve the relevant run information.
If the team chooses ClickUp because work ownership and task updates are the main concern, a ClickUp audit and workflow review may help assess workspace configuration. It is not a CPM validation service.
For ClickUp workspace implementation, ClickUp setup and automations concerns workspace setup and automation, not a verified connection to P6 or another specialist scheduling product. Confirm any required data exchange separately before making it part of the buying decision.
Preserve run-level evidence for later review
Treat each schedule calculation as a distinct run. A schedule-run record represents one project schedule version, data date, calculation configuration, and run identifier. A task observation represents one task’s result in that run. Do not combine run metadata, task float, and project-level summary metrics into a single record.
The following is an illustrative, organization-owned review-register design. It is not a vendor schema or a promised product export. Populate it through a product workflow that you have verified:
{
"schedule_run": {
"project_id": "PRJ-042",
"schedule_version_id": "update-07",
"data_date": "2026-10-23",
"calculation_method": "longest_path",
"progress_logic_method": "retained_logic",
"schedule_run_id": "RUN-20261023-01"
},
"task_observation": {
"schedule_run_id": "RUN-20261023-01",
"project_id": "PRJ-042",
"schedule_version_id": "update-07",
"task_id": "ACT-1840",
"total_float": 0,
"free_float": null,
"critical_flag": true,
"float_path_number": null
}
}The task observation grain is one task in one schedule run. The run identifier is therefore part of the observation, even when the project and task are unchanged across recalculations. Null means a value was not supplied or is not available; it is not the same as zero float.
A proposed task-observation key could be project_id + schedule_version_id + task_id + schedule_run_id. If multiple writers can create records, enforce that uniqueness in the database and use a transactional upsert where supported. A lookup followed by create is not safe against concurrent runs. If path membership is stored separately, include the run identifier, path method, endpoint, and path number in that row’s key. Citation records should likewise be one row per source citation, not one row per article or research run.
Simple deterministic checks should handle required-field validation, allowed calculation-method values, missing metadata, and exception flags. AI is not needed to calculate criticality or float. If an organization uses AI to draft an explanation of an already verified exception, restrict it to supplied facts and source record IDs. Validate the output structure, reject unsupported facts or recommendations, and require scheduler approval before the text enters a review pack. The reviewed documentation does not establish a shared API, export contract, or common field mapping across these products.
Save the schedule copy, calculation configuration, observations, and scheduler’s disposition together. Measure whether reviewers can reproduce the reported result from the retained inputs and whether material exceptions are resolved before the schedule is used externally.
Frequently asked questions about critical path software
Is a red critical-path bar the same as zero float?
Not universally. Check the product’s criticality rule and the task’s displayed float. A threshold-based critical flag and a longest-path result are different ways to identify critical activities.
What is the difference between total float and free float?
Total float concerns how much an activity or sequence can slip before affecting the project finish. Free float concerns how much it can slip before affecting a connected successor. Record the fields separately.
Can a project have more than one critical path?
Yes. Multiple paths may have the same or similar influence on the finish. The product’s method and configuration determine how they are identified and displayed. P6 documents configurable multiple float paths separately from its critical-activity setting.
Can a hard constraint create a misleading path?
Yes. Planera’s documentation warns that hard constraints can force activities onto or off the critical path and can create artificial paths that do not follow logical relationships. Inspect both the constraint and the relationship-driven chain.
Does Project Online’s retirement mean Microsoft Project desktop is retiring?
No. Microsoft’s lifecycle notice applies to Project Online, with a retirement date of September 30, 2026. It does not automatically apply to Project desktop, Project Server, or Planner.
Final selection checklist
Before approving a product, run the same representative schedule through the candidate tools and retain the evidence. Select the product only after you can answer these questions:
- What rule identifies critical activities in the edition or plan being evaluated?
- Can the scheduler inspect total float, free float, path membership, constraints, and relationship logic separately?
- Does the schedule recalculate when remaining duration, actuals, constraints, calendars, or relationships change?
- Can the team detect missing links, time gaps, open ends, and constraint-driven results?
- Can each run be identified by schedule version, data date, calculation settings, and run ID?
- Who approves results used for commitments, recovery plans, contractual reporting, or claims?
A trustworthy choice is not the tool that produces the most attractive critical-path color. It is the tool and operating process that let the responsible scheduler explain the result, reproduce it later, and identify when the network no longer represents the work.