Skip to content

Migrating from Microsoft Project Online

Microsoft retires Project Online on September 30, 2026. New PWA sites have been blocked since April 1, 2026, and after the retirement date the projects and data in a tenant are no longer reachable. Microsoft’s own successor, Planner Premium, is built for task coordination rather than scheduling: it has no critical path, no resource leveling, and no earned value, and it caps a plan at 3,000 tasks.

This page is for the PM or PMO deciding what to do with those plans. It is deliberately symmetrical — it lists what TruePPM loses against Project Online as plainly as what it keeps — because a migration chosen on an advertisement gets reversed six months later.

Every migration guide published this year says the same thing, and it is right: before you compare tools, list what your organization actually uses. Three lists matter:

  1. Active plans, by depth. How many carry real dependencies and constraints, and how many are flat task lists with dates typed in. The second kind imports into anything. The first kind is what this page is about.
  2. Reports and dashboards people depend on. Who receives a status report from Project Online today, how often, and what is on it. This is where TruePPM is weakest right now — see What you will miss below.
  3. Integrations. Power BI, Teams, SharePoint task lists. TruePPM has a REST API, webhooks, and a read-only MCP server; it has no Power BI connector.

Export each plan from Project Online as MSPDI XML (Project desktop: Save As → XML). TruePPM’s MS Project import reads it and builds a CPM-schedulable project:

  • The WBS, summary tasks, and milestones, with outline structure intact.
  • All four dependency types (FS, SS, FF, SF) with lead and lag on every link.
  • Durations in working days, and the plan’s base calendar with its non-working exceptions.
  • Start-no-earlier-than constraints (SNET), applied as the engine’s floor.
  • Actual start and finish dates, so an in-flight plan keeps its progress.
  • Three-point estimates, if your plan carries them in the standard fields, which feed Monte Carlo directly — Project Online never had this.
  • Assignees, matched or created by name. Not their rates, not their capacity.

The import lists what it dropped, per field family, so you can check the receipt against your inventory rather than discovering a gap in a review meeting.

What degrades on import, and how you will know

Section titled “What degrades on import, and how you will know”
In your planWhat happensHow it surfaces
Must Finish On, Finish No Later Than, Must Start On (the ceiling half)Dropped. TruePPM honors one constraint type today, start-no-earlier-than.A row-level warning naming each task and the MS Project constraint name it carried
As Late As Possible, Start No Later Than, Finish No Earlier ThanDropped, same reasonSame warning
DeadlineDropped. There is no deadline field yetListed in the import receipt
BaselinesReported, not imported. Capture a TruePPM baseline once the schedule is inListed in the import receipt
Work, Cost, resource ratesDropped. TruePPM schedules on duration and dependencies and has no cost modelListed in the import receipt
Sub-day durationsRounded to whole working daysSilently — check any task under one day
Resource leveling resultsNot reproduced. The engine computes the schedule and does not level itThe imported dates are the pre-leveling network dates

The constraint rows are the ones that matter for a contractual plan. The direction TruePPM intends to take — honoring FNLT, MSO, and MFO alongside SNET, with a separate soft deadline and negative float against it — is recorded in ADR-1049, proposed and not yet accepted. Until it ships, a plan that leans on must-finish-on dates cannot be expressed here, and this page will say so until it can.

Read What TruePPM doesn’t do yet in full before you decide. For a Project Online migrator the four that bite are:

  • No resource leveling. Project Online levels; TruePPM computes a schedule that assumes everyone can do everything at once. Leveling within one program is planned for 0.6.
  • No status report. A project has a health value and nothing a sponsor can read. The project status update — RAG, narrative, and the computed facts behind it, with a one-page PDF — is planned for 0.5. If your PMO’s weekly report is the thing you are really migrating, wait for it or budget for producing it by hand until then.
  • No cost or earned value. No rates, no budgets, no CPI. Planned for 0.8 to 1.0.
  • One constraint type. See the table above.
  • Monte Carlo in the core. P50 / P80 / P95 finish dates and per-task sensitivity, computed by the same engine that computes the deterministic dates. Project Online never had this; the commercial add-ons that provide it are licensed separately.
  • The board and the schedule are one object. A team that works in sprints moves cards; the PM’s critical path moves with them. No parallel plan.
  • Your infrastructure, your data. Apache 2.0, self-hosted, no per-seat license, OIDC login against your own identity provider.
  • An API that is the product. Everything the web UI does is a REST or WebSocket call you can make yourself, plus a read-only MCP server for asking questions of the live schedule.
  1. Pick one real plan with dependencies and at least one hard finish date. Export it as MSPDI XML.
  2. Import it into the hosted demo or your own install via Bring your existing plan in. Read the import receipt against your inventory.
  3. Compare the finish date TruePPM computes with the one Project Online showed. A difference is usually a dropped constraint or a leveling delay; the receipt tells you which.
  4. Run Monte Carlo on it and look at the P80. This is the number you did not have before.
  5. Try to produce the report your sponsor gets today. If you cannot, that is the honest reason to wait for 0.5, and this page will change when it lands.