Leaving a cloud provider is usually discussed as a purchasing decision: which European alternative exists, what it costs, when the contract ends. Our reading of the Eurostat and ACM data is different. The binding constraint is people. Most Dutch companies buy cloud services, a much smaller group employs anyone with ICT specialist skills, and the firms that did manage a switch mostly brought in outside capacity to do it.
Deep dependence, thin staffing
Buying email in the cloud is not the same as running a production database there. On that deeper layer, hosting databases and files with a provider, Dutch firms sit near the top of the European Union, far above the European average. The Netherlands is not unusual in using cloud. It is unusual in how much of its core company data it has placed there.
Staffing has not kept pace. Across firms of all sizes, cloud use is widespread while in-house ICT specialists are the exception. The gap is narrow at large employers, which mostly have their own specialists. It is widest at small firms, which buy cloud almost as readily but rarely employ anyone who could plan a migration.
Why hiring does not close the gap
The Netherlands has a large ICT workforce by European standards. The difficulty is in hiring. Dutch firms try to recruit ICT specialists more often than their European peers, and most of those that try are left with a vacancy they cannot fill. The reason employers give most often is not pay. It is that nobody applies. Missing qualifications and missing experience follow. Salary expectations come last.
That matters for anyone planning an exit. A migration off a hosted database or a development platform is an engineering project with a real duration. When the competition regulator ACM asked business cloud users about switching, a sizeable minority of attempts had failed, and among firms that succeeded, most had used an intermediary. Where a move happened, the capacity was more often bought in than supplied from inside.
What this does not show
The switching survey ran in the winter of 2024 and 2025, before geopolitics gave many firms a reason to want to leave. Demand for migrations may have risen since, and the data cannot say by how much. The ICT labour market has also eased since 2022: vacancy rates fell and recruitment difficulty dropped somewhat, so hiring is less hard than at the peak. Dutch cloud adoption dipped slightly in the latest wave, which may be noise. Finally, these are national survey figures. They do not show how dependent any single company is, and the official ICT staffing data stops at 2024. If a later wave showed recruiting firms filling ICT roles without difficulty, our reading would weaken.
What you can do
- Cost an exit in people, not licences. Write down which roles the migration needs, for how long, and who would fill them today.
- Decide early whether you hire or contract that capacity. The data supports one or the other, not a plan that assumes the current team will absorb it.
- Start recruiting before you sign with a new provider. The vacancy, not the contract, sets your timeline.
- Approach people directly. The most cited problem is that nobody applies, so a better advertisement or a higher offer alone will not fill the role.
- If you are a small firm with deep cloud use, treat this as urgent. That is where dependence and in-house knowledge sit furthest apart.
For a view of this field, see cloud, DevOps and infrastructure.