What it actually costs to migrate from MS Project Online in 2026

Small but specific search interest exists for this question — people are actively researching migration costs ahead of the September 2026 Project Online retirement. Until now, no direct answer existed online. This post is that answer.

Most migration-cost guides are vendor pitches. They either minimize the cost ("just import and you're done!") to make the destination tool look free, or inflate it to justify a consulting engagement. We don't do either. We built Onplana to be the best Microsoft Project alternative for schedule-driven PMOs and we think the numbers speak for themselves — but that only works if the numbers are honest.

This post walks through the six cost categories that make up a full 3-year migration budget: license delta, data-migration labor, training and change management, parallel operation, integration rework, and cleanup. For each category we show the underlying math, the ranges our cost calculator uses, and where the cost varies based on the choices you make. At the end there's a free calculator that produces a specific number for your organization — but the categories matter more than any specific estimate, because what you need from a CFO is a defensible line-item budget, not a total.

The 6 cost categories most organizations forget to budget

Migration budgets that come in over tend to break down the same way: labor was underestimated, integration rework wasn't budgeted at all, or someone assumed training was "self-service" and then had to bring in an instructor three weeks into go-live. Below is the full category set we model, in the order they typically hit the budget.

License delta — the tip of the iceberg

The license delta is the 3-year cost of continuing to pay Microsoft if nothing changes, set against what you'll pay the new tool's vendor for the same period. Our calculator uses your current annual spend × 3 × (0.8 to 1.2) as the range — ±20% captures Microsoft price-tier changes and seat-count drift. For a PMO paying $50,000/year across 50 Project Online P3 seats, the license delta range is $120,000-$180,000.

The important framing: migration cost displaces license cost, not adds to it. If Onplana's 3-year license comes in $40,000 below what you'd pay Microsoft, that's $40,000 of migration cost you're "pre-paying" with savings you'd otherwise bank. Most vendor cost comparisons hide this by presenting migration as pure additional spend. Read Onplana's pricing at /pricing and compare directly against your Microsoft E-plan pricing before accepting any vendor's cost narrative.

License delta is also where tool selection matters most. The cheapest reasonable destination and the most expensive differ by 3-4x on license. A 50-seat PMO moving to a budget tool might spend $30,000/year in license; the same PMO moving to an enterprise suite with heavyweight governance modules might spend $120,000/year. The 3-year gap is six figures. It's worth spending real time on the tool-selection step before committing — our detailed Onplana vs Microsoft Project comparison walks through the per-seat math side by side.

Data migration labor — the big one

Data migration is almost always the biggest single cost category. Our calculator uses 3-12 hours per active project and 2-6 hours per historical schedule at a blended $100/hour rate. The range is wide because schedule quality varies enormously — a clean dependency-driven schedule with no custom fields imports in 2-3 hours, a resource-loaded cost plan with 40 custom field values and hard constraints needs a day of manual reconciliation.

For a typical mid-size PMO with 60 active projects and 200 historical schedules, that's $58,000-$192,000 in data-migration labor. The spread is dominated by two variables: your destination tool's native .mpp import quality, and how much cleanup you do per schedule.

Tools with native .mpp import (Onplana's import parses tasks, dependencies, resources, custom fields, and baselines in one pass) land near the low end. Tools that require you to export each schedule to CSV or XML and script a custom importer land near the high end. The schedule health check tool gives you a per-file indicator of which schedules carry cleanup debt — if 40% of your schedules show critical issues, plan for the high end of the range.

The 3-year framing matters here because data migration isn't always one shot. Historical schedules get migrated in a second wave 6-12 months after go-live as the business decides which ones it actually needs. Budget for this phased reality, not a single push.

Training and change management

Training is 4-12 hours per PM at a blended rate, with a geo multiplier (1.0-1.5) that captures the overhead of coordinating training across multiple regions and time zones. A 50-PM single-location team budgets $20,000-$60,000. A 50-PM fully-distributed team budgets $30,000-$90,000.

Change management — the non-training component — is the part most organizations forget. This includes: executive communications to explain why the change is happening, manager-level reinforcement in team meetings for the first 30 days post-cutover, early-adopter champions who get trained first and help others, and the internal knowledge base or wiki that captures tool-specific patterns your PMs will reference for the next two years. Budget 20-30% on top of formal training hours for change management.

The change-management cost shows up differently in different organizations. Some PMOs over-invest here (expensive external consultants running workshop series) and some under-invest (one lunch-and-learn and a PDF handout). The right level is proportional to PMO maturity: teams that already have strong internal documentation norms need less; teams where institutional knowledge lives in individual heads need more.

Parallel operation period

During the cut-over you'll run both tools concurrently — the old PWA site for in-flight schedules that aren't ready to move, the new tool for schedules that have moved. Our calculator uses 3 months as the default parallel-operation period with a 1.0-1.5× multiplier on quarterly current spend, plus 2-4 hours per PM per month of dual-tooling overhead.

For a 50-PM PMO with $50,000/year current spend, that's roughly $12,500-$18,750 in dual license, plus $30,000-$60,000 in PM overhead time. The PM overhead is almost always the larger piece and the most frequently forgotten.

The parallel period is shorter if your destination tool has good .mpp ingestion — you can move schedules in batches rather than rebuilding them, which means individual PMs don't run dual-toolset for as long. Phased migrations extend the parallel period in exchange for smoother risk; big-bang migrations compress it but spike PM overhead into a short window.

Integration rework — the scope bomb

Integration rework is the category most likely to scope-bomb your migration. Our calculator uses $5,000-$20,000 per confirmed integration and $2,000-$10,000 per "don't know" integration (half-rate, reflecting that some unknowns turn out to be non-integrations and others turn out to be complex).

Based on patterns we've seen in migration consulting, three integration categories drive most of the cost: Power BI reports consuming the OData feed, SharePoint-driven governance workflows, and custom-field-dependent exports to downstream systems (SAP, Workday, or homegrown tools). A single Power BI report rebuild typically runs 2-5 person-days once you account for column mapping, authentication, visual recreation, and stakeholder acceptance.

The uncertainty here is why the calculator's sensitivity analysis often flags integrations as the #1 or #2 driver of total cost variance. If you answered "don't know" to all three integration-discovery questions, your estimate's range is dominated by integration uncertainty. A one-day discovery pass with IT typically tightens the range by 30-50%.

Cleanup work — the hidden cost

Cleanup is the per-schedule work to fix broken dependencies, resolve orphaned resources, retire obsolete custom fields, and clean up the detritus accumulated over years of tool use. Our calculator uses $30-$100 per historical schedule plus 1-3 hours per PM for review time.

For a 200-historical-schedule, 50-PM PMO, that's $11,000-$35,000. Sounds small. The reason it's not larger is that most cleanup is batched — scripts run across the historical estate, a small team of reviewers spot-checks the output, and the per-schedule cost is mostly script-and-review, not hand-editing.

If you skip cleanup entirely, you carry forward problems to the new tool and pay for them later. Most schedules carry hidden quality issues that migration will preserve but won't fix. The cleanup line item is where you decide whether to fix them now or pay for them downstream.

Cost by organization size

Order-of-magnitude 3-year ranges across organization sizes:

• Small PMO (under 10 PMs, under 30 active projects): $30,000-$120,000. Dominated by training and license delta; data-migration and integration rework are small in absolute terms.
• Mid-size PMO (10-50 PMs, 30-120 active projects): $130,000-$450,000. Balanced across all six categories; data migration and integration rework often tie for largest single line item.
• Enterprise PMO (50+ PMs, 120+ active projects): $450,000-$1.8M. Integration rework dominates — multiple heavily-customized downstream systems, SAP/Workday connections, regulatory reporting dependencies.

These are rough ranges; a specific estimate for your org requires running the actual inputs. An enterprise PMO with a clean integration surface can land in the mid-size range; a mid-size PMO with a spaghetti of Power Automate flows can land in enterprise territory. Sensitivity to integration count is the dominant factor across all three segments.

How to get your specific number

We built a free migration cost calculator that runs the six-category math against your specific inputs. Enter your PM count, active-project count, historical-project count, current annual spend, geographic distribution, and integration posture, and it produces a low/high 3-year range with per-category breakdown, sensitivity analysis identifying your top three cost drivers, and an Onplana comparison showing where migration cost reduces versus a generic destination.

The calculator output is designed to be pasted directly into a business case. The per-category breakdown maps to budget line items; the sensitivity analysis gives you three specific "if X changes, cost moves by Y" statements that a CFO can sanity-check against real data.

How to build your internal business case

A credible internal business case has four parts, in order. First, the fixed deadline: Project Online retires September 30, 2026, and doing nothing is not an option. Second, the 3-year cost range, pulled from the calculator for your specific inputs. Third, the sensitivity analysis identifying which inputs drive the range — this gives your executive team the levers they can pull to reduce cost (archiving historical schedules, sharpening integration discovery, choosing a destination with good .mpp import). Fourth, a tool-selection recommendation backed by the first three.

The most common mistake in internal business cases is leading with the tool recommendation rather than the cost math. CFOs read cost math; they discount tool recommendations from anyone who obviously has a preferred tool. Lead with the numbers.

Conclusion

A defensible migration budget is a line-item budget, not a total. The six categories in this post are what a CFO actually needs to see: license delta, data-migration labor, training and change management, parallel operation, integration rework, and cleanup. Ranges matter more than point estimates — a number with a range lets stakeholders probe assumptions. A total without a range gets pushed back until someone produces one. The free calculator gives you the full six-category breakdown in under two minutes; once you have it, the business case writes itself.

Frequently asked questions

Is the cost the same regardless of destination tool? No. License delta varies by tool (Onplana's per-seat is lower than most enterprise PM suites). Data-migration labor varies by whether the destination has native .mpp import. Integration rework varies by how close the destination's API surface is to what your existing connectors expect. The cheapest and most expensive reasonable destinations for the same org often differ by 2-3x on total 3-year migration cost.

How much should I budget for contingency? 15-25% on top of the high-end estimate is reasonable for a mid-size PMO. Bias toward 25% if you answered "don't know" on any integration-discovery question. 15% is defensible only if your integration surfaces are well-documented. Contingency is for genuinely unknown work — don't use it to pad known-work estimates.

What's the cost of NOT migrating? After September 30, 2026, Project Online stops serving. There is no "extend the subscription" option. The cost of not migrating is the cost of losing access to every active schedule, report, integration, and resource calendar simultaneously. Organizations that plan to "deal with it later" end up paying premium emergency-migration rates in Q3 2026 that typically run 2-4x a planned migration.

Can I phase the migration to spread cost? Yes, and most organizations with 100+ projects do. Common pattern: move active schedules first (cheapest per schedule, highest value), archive historical schedules to read-only storage, phase custom-field translation over three to four sprints. Phased migrations have slightly higher total cost because parallel-operation overhead lasts longer, but they spread labor and risk.

What does "integration rework" actually include? Any external system reading from or writing to Project Online that needs repointing: Power BI reports on the OData feed (2-5 person-days each to rebuild), SharePoint-driven governance workflows (rebuild not migrate, 1-3 person-days each), and custom-field-dependent exports to Excel, SAP, Workday, or homegrown systems (highly variable). Our calculator uses $5,000-$20,000 per confirmed integration based on patterns we've seen in migration consulting.

How much of this is one-time vs recurring? Most is one-time. License delta is the only recurring cost. Data migration, training, parallel operation, integration rework, and cleanup are all one-time. The distinction matters for finance — capitalize one-time migration work as a project expense, treat license delta as an opex shift.

Should I include my own PM time in the cost? Yes. The calculator's PM-time components (2-4 hours per PM per month during parallel operation, plus training hours) are there for exactly this reason. The most common underestimate in migration projects is the PM leadership team's time. If you exclude PM time, you'll finish the migration but miss other commitments that quarter.