ERP Implementation Statistics & Success Rates

Depending on which study you read, between 55% and 75% of ERP implementations fail to meet their original objectives, and most of the rest still arrive late and over budget. The most rigorous recent audit of 640 finished projects found that only 32% came in within budget and only 27% hit their planned go-live date. The headline numbers are real enough to take seriously, but they are far more useful once you understand what each study is actually counting, how outcomes vary by platform and industry, and which interventions the data shows move the odds.

This is the data view. For a deeper treatment of why these failures happen — the organizational and technical root causes behind the numbers below — our guide to ERP failure causes walks through each driver in depth, and our ERP ROI framework covers how to model the financial outcomes these statistics describe.

Why the headline failure rate varies so much

The first thing to understand about ERP implementation statistics is that "failure" is not a standardized term, so different studies measure different things and arrive at deceptively similar-looking numbers.

Gartner puts the failure rate at 55% to 75%, capturing projects that either failed outright or "delivered significantly below expectations" — a deliberately broad definition that sweeps in everything from abandoned projects to systems that went live but were never properly adopted.
McKinsey reports that 70% of large-scale technology transformations fail to meet their goals, but this figure draws on broader digital-transformation research (not ERP alone) and measures failure against the original business case, covering ROI targets, productivity gains, and cycle-time reductions.
IDC found that 75% of ERP deployments experience significant delays, a narrower metric that counts only schedule slippage regardless of whether the project ultimately delivered value.
The Standish Group's CHAOS Report consistently finds that 15% to 20% of large IT projects are cancelled before completion, with sunk cost at cancellation typically 60% to 80% of the original budget.

When you layer these definitions, you can see why the numbers look alike but aren't quite the same: they are measuring different failure modes. Most ERP projects experience at least one of them, which is why almost every practitioner will tell you the broad percentages are directionally correct even if the precise figure should be held loosely. (Net Fusion Technology)

One forward-looking forecast worth tracking

Gartner's most actionable number is a prediction, not a retrospective. In 2024 it forecast that by 2027, more than 70% of recently implemented ERP initiatives will fail to fully meet their original business case goals, and as many as 25% will fail catastrophically. That forecast matters because it tells you the structural problems behind the failure rate have not gone away with cloud deployment — they are expected to persist. (Gartner)

What the largest studies actually found

Rather than quote a single failure rate, it is more honest to lay the major datasets side by side. Each was built differently, and together they describe the same underlying reality.

Gartner (2024) — Sample: Industry forecast · Metric: Fail to meet business case by 2027 · Finding: >70%, with ~25% catastrophic
McKinsey Digital — Sample: Large-scale transformations · Metric: Fail to meet goals · Finding: ~70%
IDC (2024) — Sample: ERP deployments · Metric: Significant schedule delay · Finding: 75%
Panorama Consulting, 2026 ERP Report — Sample: 170 organizations (Jan 2025–Jan 2026) · Metric: Projects over budget · Finding: 30%
Panorama Consulting, 2026 ERP Report — Sample: Same · Metric: Projects over schedule · Finding: ~25%
Panorama Consulting, 2024 ERP Report — Sample: 131 organizations (Aug 2022–Dec 2023) · Metric: Over budget · Finding: ~47% (prior report)
DataVirtualizer 2026 audit — Sample: 640 finished projects (Jan 2024–Jun 2025) · Metric: Within original budget · Finding: 32%
DataVirtualizer 2026 audit — Sample: Same · Metric: Met original go-live date · Finding: 27%

Two things stand out. First, the more granular and recent the audit, the clearer the picture: in the 640-project dataset, fully 68% of ERP implementations exceeded their originally approved budget, with a median overrun of 47% above the initial budget. A concerning 19% of projects exceeded budget by more than 100%, effectively doubling the original investment. Second, the long-running Panorama survey shows real improvement over time — the over-budget rate has fallen from around 47% in its 2023 report to 30% in the 2026 report — even as the underlying risk has not disappeared. (DataVirtualizer, Panorama Consulting via ERP-Software.org)

Budget overruns: what projects actually cost

The budget data is the most consistently tracked dimension, and it is where the financial pain concentrates.

The 640-project audit reported a median project budget of $1.8 million across organizations ranging from 50-employee mid-market firms to 25,000-employee enterprises. In that dataset, 34% of over-budget projects exceeded budget by 20–50%, while the long tail of 19% more than doubled the spend. (DataVirtualizer)

Panorama Consulting's survey work gives the most useful benchmarks for planning, because it breaks cost down by company profile:

2024 ERP Report (midsize companies, median $200.5M revenue, 750 employees): median project cost $450,000.
2023 ERP Report (more enterprise-heavy sample, median $1.5B revenue, 6,500 employees): median project cost $625,000, median timeline 15.5 months.
Software Path's 2022 ERP Report (1,384 real selection projects, mostly US/UK): average budget of $9,000 per user, up from $8,295 per user in 2021.

(Panorama Consulting via ERP-Software.org, Software Path)

What drives the overruns

Panorama's 2026 report is unusually specific about why budgets blow out, and the answers are revealing because they are mostly controllable. Among over-budget projects, the most common causes were:

Unexpectedly needed additional technology — 54.9%
Expansion of the original project scope — 51.0%
Technical issues — 43.1%
Organizational issues — 39.2%
Project staffing underestimated in the initial budget — 35.3%

Notice that three of the top five causes are planning and governance problems, not technology problems. The budget rarely blows up because the software is broken; it blows up because the scope and the integration footprint were never fully counted. (Panorama Consulting via ERP-Software.org)

Timeline overruns: the schedule data

If budgets are bad, schedules are worse. In the 640-project audit, only 27% of projects met their original go-live date, and the median schedule overrun was 5.3 months beyond the planned completion — a 41% extension of the originally planned timeline.

The delays are not evenly distributed across platforms:

SAP S/4HANA — 7.1 months
Microsoft Dynamics 365 — ~5 months (mid-pack)
Workday — ~4.5 months
Oracle NetSuite — 3.8 months

The gap reflects the relative complexity difference between on-premises-style enterprise ERP and cloud-native mid-market deployments. (DataVirtualizer)

Timelines are actually getting shorter

There is genuinely good news in the trend data. Panorama Consulting's median project timeline has fallen from 15.5 months (2024 report) to 9 months (2026 report) — a decline of more than 40% in two years. And in its 2026 sample, only about a quarter of companies exceeded their planned timeline, down meaningfully from earlier years. The most common reason for the remaining schedule overruns was organizational — governance, resistance to change, and process redesign — rather than technical execution. (Panorama Consulting via ERP-Software.org)

In other words: cloud deployment, better migration tooling, and more agile methodologies are compressing the calendar. What they have not yet fixed is the human and governance layer.

Scope and benefits realization: the ROI gap

This is where the failure statistics get most uncomfortable, because a project can be delivered on time and on budget and still fail commercially.

In the 640-project audit, 71% of projects delivered the core functionality specified in the original scope — but only 44% delivered all planned functionality, including integrations, custom reports, and workflow automations. The most commonly descoped items were:

Advanced reporting and analytics (cut from 38% of projects)
Third-party integrations (cut from 34%)
Workflow automation (cut from 29%)

(DataVirtualizer)

The benefits-realization data is where the money story lives. The audit found that organizations in the bottom quartile of user adoption realized only 41% of projected business benefits within the first year. This is the implementation that declares go-live success while the business quietly keeps running workarounds in the legacy system — the failure mode McKinsey's 70% figure largely captures.

There is a hopeful data point here too: Panorama's 2023 report found that 83% of organizations that performed an ROI analysis prior to their ERP project, and had been live for at least a year, said their project met their ROI expectations. Upfront financial discipline correlates strongly with downstream value. For the mechanics of building that discipline, see our ERP ROI framework, which covers how to model benefits, total cost of ownership, and payback honestly.

How outcomes differ by platform

Platform choice has a measurable, sometimes dramatic effect on cost and timeline outcomes. The 640-project audit did a focused comparison of SAP S/4HANA versus Oracle NetSuite for organizations with 200–2,000 employees:

Median total cost — SAP S/4HANA: $2.4 million · Oracle NetSuite: $680,000
Median timeline — SAP S/4HANA: 16.5 months · Oracle NetSuite: 8.2 months
% completed within budget — SAP S/4HANA: 24% · Oracle NetSuite: 41%

NetSuite's superior budget predictability was attributed to its standardized implementation methodology (SuiteSuccess), lower customization requirements for mid-market use cases, and a cloud-native architecture that eliminated infrastructure-provisioning variables. SAP projects delivered broader functional depth — particularly in manufacturing, supply chain, and multi-entity consolidation — but at roughly 3.5 times the median cost and twice the timeline. (DataVirtualizer)

The platform distribution in that same dataset was SAP S/4HANA (28% of projects), Oracle NetSuite (24%), Microsoft Dynamics 365 (22%), Workday (14%), and other platforms including Sage Intacct, Acumatica, and Epicor (12%).

How outcomes differ by industry and company size

Failure risk is not uniform. Industry and business model compound it significantly.

A 2026 analysis drawing on Panorama Consulting's ERP Report and a dataset of more than 2,400 discrete-manufacturing implementations found that discrete manufacturing records the highest failure severity of any segment studied:

Implementation failure rate — Industry average: 68% · Discrete manufacturing: 73%
Average budget overrun — Industry average: 189% · Discrete manufacturing: 215%
Timeline extension — Industry average: 25% · Discrete manufacturing: 30%
Objective achievement — Industry average: 32% · Discrete manufacturing: 27%

(Godlan)

The same analysis scored business-model complexity on a 100-point index, and the escalation is steep: make-to-stock sits at 65/100, make-to-order at 78, configure-to-order at 85, and engineer-to-order at 92. The 27-point gap between make-to-stock and engineer-to-order reflects fundamentally different implementation requirements — product configurator expertise, CAD/PLM integration, dynamic pricing — that generalist teams routinely underestimate during scoping.

Aggregated industry data tells a consistent story across sectors: public-sector ERP projects fail most often (~78%), followed by manufacturing (~72%), retail (~69%), finance (~64%), and healthcare (~50%). Larger companies also report worse outcomes — only about 11% of larger firms describe their rollout as smooth, versus just under 22% of smaller companies, according to the Trovarit ERP in der Praxis study. (Gitnux, ERP-Software.org)

Cloud vs. on-premise: does deployment model change the odds?

A reasonable question is whether the move to cloud ERP has actually moved the failure statistics. The honest answer is that it has improved some dimensions — timelines, infrastructure risk, budget predictability — without fixing the organizational ones.

Aggregated deployment-model data suggests cloud ERP fails less often than on-premise, but not by a wide margin: cloud ERP projects fail at roughly 48%, on-premise at 63%, and hybrid models at 55%. The improvement is real and consistent with the timeline compression noted earlier — cloud-native architectures remove a whole class of infrastructure-provisioning risk — but cloud projects are not immune. A meaningful share still miss their objectives for the same governance and adoption reasons that dog every deployment model. (Gitnux)

The more important deployment statistic is about where the money goes over the system's life. Across studies, roughly 70% of total ERP cost is incurred after go-live in support, maintenance, optimization, and upgrades, and hidden costs account for an estimated 40–50% of total implementation expenses. (Gitnux) That ratio is the real argument for treating implementation as the start of a multi-year program rather than a project with a finish line: the decisions you make during implementation — how much you customized, how cleanly you migrated data, how well you trained users — determine the size of the recurring cost stream for years afterward. This is also why benefits realization is the decisive metric, not go-live: a project that ships on time but lands heavy customization debt will pay for that decision every quarter until it is retired.

The deeper point about root causes

Notice that across every dimension — budget, timeline, scope, deployment model — the failure drivers the data keeps surfacing are organizational, not technical. Insufficient requirements, poor data quality, underfunded change management, over-customization, weak governance, and inexperienced partners account for the overwhelming majority of overruns. The technology is rarely the binding constraint. If you want to understand any one of those drivers in depth rather than as a statistic, the ERP failure causes guide breaks down each root cause and how to prevent it.

What separates the projects that beat the odds

The single most useful finding across all of this research is that the organizations that escape the failure statistics do not do so by being luckier. They invest in specific, measurable ways, and the data quantifies the payoff of each.

Requirements and design discipline

Insufficient requirements definition was the leading root cause in the 640-project audit, identified in 72% of over-budget projects. Organizations that invested less than 8% of total project budget in the requirements and design phase experienced 3.2 times higher rates of budget overrun than organizations that invested 12–15%. The recommended allocation is 12–15% of total project budget in requirements and design, including comprehensive business-process documentation and data-quality assessment.

Data migration, started early

Data-migration complexity affected 64% of delayed projects, and organizations underestimated the effort by a median of 180% — the actual work was nearly three times the planned allocation. Legacy data-quality issues (duplicate records, inconsistent formatting, missing required fields) accounted for 60% of data-migration delays. The organizations that conducted comprehensive data-quality assessments before implementation planning experienced 52% fewer data-migration delays. The implication is blunt: start data profiling in sprint one, not month four.

Change management, funded properly

This is the most consistently replicated finding in the entire implementation literature, and the numbers are large. Organizations that allocated less than 5% of project budget to organizational change management reported median user-adoption rates of 54%, compared to 83% for organizations investing 10–15%. Prosci's research across more than 2,000 change-management studies finds that projects with excellent change management are six times more likely to meet their objectives than projects with poor change management. The recommendation is a dedicated change-management budget of 10–15% of total project cost, started in month one — not a training event scheduled two weeks before go-live.

Customization, kept minimal

Customization scope creep was identified in 47% of over-budget projects, and it had the strongest single correlation with cost overrun in the entire dataset (r = 0.74). Organizations that modified more than 15% of standard ERP workflows experienced median cost overruns of 72%, versus 28% for organizations that kept customization below 5% of standard workflows. The target is explicit: less than 5% modification of standard workflows, achieved by adapting business processes to the software rather than the reverse.

Integration, counted honestly

Integration complexity accounted for overruns in 51% of projects. The median ERP implementation required integration with 14 external systems (CRM, HRIS, banking, e-commerce, warehouse management, and similar), and each integration added a median of $32,000 to project costs and 3.2 weeks to the timeline. Organizations that underestimated their integration count by more than 30% experienced 2.4 times the rate of budget overruns.

Implementation partner selection

Inadequate systems-integrator selection contributed to 39% of failed projects. Organizations that selected their integrator based primarily on price experienced 2.8 times higher failure rates than those that weighted platform experience and reference quality above price. Integrators with fewer than five completed implementations on the selected platform showed 3.1 times higher rates of significant budget overrun. The leverage here is enormous: partner expertise is the one decision that either amplifies or mitigates every other risk on the list. (All figures in this section: DataVirtualizer. For how to structure that partnership, our ERP implementation services cover scoping, governance, and delivery.)

How to actually use these statistics

The aggregate failure rate is less useful than understanding the distribution of failure modes and their respective leverage points. Translated into planning action, the data points to a concrete set of risk-adjusted practices:

Build a contingency reserve of 25–35% above base budget estimates. The median overrun across studies is 27–47%, so a 10% contingency is not realism — it is denial.
Plan a timeline buffer of roughly 30% above integrator-projected durations. The dataset shows systematic optimism bias in vendor estimates.
Front-load requirements and data-quality work. Spend 12–15% of the budget before build starts, and begin data profiling immediately.
Budget change management as a workstream, not an event. Allocate 10–15% of project cost, starting at kickoff.
Cap customization at 5% of standard workflows, and challenge every change request against that ceiling.
Count integrations explicitly during discovery, and treat any undercount as a budget risk, not a detail.

Each of these practices is backed by a multiplier in the data — typically two to six times better outcomes for the organizations that adopt them versus those that do not.

The bottom line

ERP implementation statistics are genuinely sobering — between roughly 55% and 75% of projects miss their objectives, most run over budget, most run late, and many never deliver the ROI that justified them. But the numbers are not a verdict on ERP as a category. They are a map of where the leverage is. The same research that produces the alarming headlines also shows, with remarkable consistency, that the organizations which invest in requirements discipline, early data work, funded change management, minimal customization, honest integration counting, and an experienced implementation partner move their odds dramatically.

The headline number is a call to attention. The breakdown — the budget, timeline, scope, and benefits-realization data broken out by platform, industry, and root cause — is where the program design lives. Treat the failure rate as a planning input, apply the interventions the data validates, and the statistics stop being a prediction of your outcome and start being a checklist of what to get right.