Cloud Exit for SMEs: Why Companies Are Bringing Their ERP Systems Back
Exploding SaaS costs and vendor lock-in: first mid-market companies are bringing ERP systems back. The business case for cloud repatriation. Read now.
For five years, the message was clear: if you don’t move to the cloud, you’ll be left behind. The Bitkom Cloud Monitor confirmed rising adoption numbers year after year. Analysts, consultants, software vendors – everyone preached the same direction. On-premise was the legacy model. Anyone still running SAP, Microsoft Dynamics, or an industry-specific ERP on their own infrastructure was either brave or behind the times.
Now the mood is shifting. Not loudly, not on conference stages. But in the meeting rooms of IT leaders. The question is no longer “When do we move to the cloud?” – it’s: “Is it still worth it?” This strategic assessment is a core topic in my ERP consulting for SMEs.
What Happened: The Cloud Bill Arrives
The first three years of a cloud migration usually feel good. The infrastructure is modern, updates roll out automatically, admin overhead decreases. But then something happens: costs start rising. Gradually, but steadily.
Typical cost drivers I see in projects:
- Annual license increases of 8-15%, with little contractual room to negotiate
- Usage-based billing that explodes as data volume or user counts grow
- Add-on costs for features that were included in the on-premise model as standard
- Egress fees – charges for exporting data out of the cloud that only become visible when you try to leave
A mid-market company with 200 ERP users can easily pay EUR 300,000-500,000 per year for a cloud ERP today. For a system they ran five years ago for significantly less on-premise. The total cost of ownership over ten years is higher than self-hosted operation in many cases – with less control.
Vendor Lock-In: Getting Out Is Harder Than Getting In
The cloud migration was technically demanding but strategically simple: the vendor had a strong interest in supporting the switch. The way back is the opposite.
Data sits in proprietary formats. Customizations, workflows, and integrations are tied to the cloud platform. Exporting to another system is possible – but expensive and labor-intensive. That’s by design.
APIs that only work in one direction. Many cloud ERPs offer excellent import interfaces. Exporting complete, consistent datasets? Far less convenient.
Contractual lock-in. Multi-year terms, minimum commitments, and cancellation periods that prevent a quick switch. If you want out of the cloud, you need to start planning 12-18 months ahead.
Cloud Repatriation: Not a Step Back, but a Calculation
When I talk about a “cloud exit,” I don’t mean: set up servers in the basement and go back to 2015. The projects I’m currently supporting are more nuanced than that.
The Hybrid Model Wins
Most companies don’t bring everything back. They bring the core ERP system back – financial accounting, inventory management, manufacturing – and run it on their own infrastructure or with a regional hosting provider. Collaboration tools, CRM, and communication stay in the cloud.
The argument is economic, not ideological. An ERP system with stable user counts and predictable data volumes runs cheaper on-premise. A collaboration tool that thrives on flexible user numbers is better off in the cloud.
The Business Case in Numbers
A concrete example from a recent project: a manufacturing company with 180 users pays EUR 420,000 per year for its cloud ERP. Repatriating to a hosted solution costs approximately EUR 280,000 one-time (migration, licenses, infrastructure). Ongoing costs afterward: EUR 160,000 per year. Break-even after 14 months.
This calculation doesn’t work in every case. But it works far more often than cloud vendors would like to admit.
Prerequisites: What to Clarify Before the Exit
A cloud exit is not an impulse purchase. Anyone planning seriously needs to answer three questions.
1. How Cleanly Can the Data Come Out?
Test early what export options your cloud vendor offers. Master data, transactional data, document archives, customization settings. Run a test export – not in two years, but now. The sooner you know where the gaps are, the better you can plan.
2. What Infrastructure Is Realistic?
An in-house server is one option, but not the only one. Managed hosting with a local provider offers control without the operational overhead. Colocation in a data center is another option. The right choice depends on your IT team and your security requirements. The BSI C5 criteria catalogue provides a solid framework for evaluating hosting providers.
3. What Does the Migration Plan Look Like?
A cloud exit is a migration project – with the same risks as the original cloud migration. Data migration, testing phases, parallel operation, training. Plan for 6-12 months. And don’t make the mistake many companies made during the cloud migration: underestimating the effort. I’ve documented the typical pitfalls of ERP migrations in detail in my article ERP Migration: 5 Mistakes That Make Projects Fail – they apply equally to a cloud exit.
What This Means for IT Strategy
The cloud exit trend is not an anti-cloud manifesto. It’s a sign that mid-market companies are maturing in how they handle cloud technology.
The first wave was emotional: cloud is modern, on-premise is old. The second wave is analytical: what’s cost-effective, what gives us control, what fits our company size?
Not every company should leave the cloud. But every company should do the math. And those who do will find, in more than a few cases, that the emperor has no clothes.
Bottom Line
The question is not cloud or no cloud. The question is: which workloads belong where? A core ERP system with stable requirements and high data volumes is often more economical on-premise or in managed hosting. Flexible, collaborative applications continue to benefit from the cloud.
If you’re approaching a contract renewal, now is the time to honestly calculate the total cost of ownership – including the costs the cloud vendor doesn’t put on page one of the proposal. The results are often surprising.
Next Step
Facing a cloud contract renewal and wondering if an exit makes sense? I run the TCO analysis and guide the migration path – vendor-neutral and outcome-driven.
→ Or read more first: DATEV Integration Audit