Under Dutch law, a service level agreement (SLA) is an ordinary contract: its commitments are binding and enforceable like any other contractual obligation. The main exception lies in the wording: a service level only becomes an enforceable obligation of result if it is written as one, and not as a target or an effort.
What distinguishes an SLA is that it tries to make performance measurable. Whether it succeeds depends on one drafting choice: is each commitment an obligation of result (resultaatsverbintenis) or an obligation of best efforts (inspanningsverbintenis)? That distinction decides what happens when the service falls short. Where availability is expressed as a defined percentage over a defined measurement period, failing to reach it is a failure to perform, and the agreed consequence follows. Where the provider undertakes to “use reasonable endeavours” to maintain availability, the customer must prove that the provider did not act as a reasonably competent provider would have done, which is a much harder case to make. The same applies to response and resolution times: a target is not an obligation unless it is written as one.
Three further elements determine whether an SLA works in practice. Service credits are usually the agreed remedy. If the contract makes them the exclusive remedy, the customer gives up its right to damages and often its right to terminate. Such a limitation is valid in principle, but under Article 6:248(2) BW it cannot be invoked where that would be unacceptable by standards of reasonableness and fairness, particularly when the failure was serious. Measurement must be defined: who measures, with which tools, over which period, and what is excluded as planned maintenance or force majeure. And the exit must be arranged in advance, because the moment a customer most needs its data returned in a usable format is the moment relations have broken down.
Table of contents
What is an SLA and what is its legal status?
An SLA sets out the service levels a provider must deliver, how they are measured and what happens if they are not met. It can be a separate contract or an annex to a main agreement, such as an IT services, cloud or outsourcing contract.
Which rules of Dutch law apply to an SLA?
Dutch law has no specific statute for SLAs. The general rules of contract law in Book 6 of the Dutch Civil Code (Burgerlijk Wetboek, BW) apply, and for most services also the rules on the contract for services (overeenkomst van opdracht) in Articles 7:400 and following BW.
Under Article 7:401 BW, a service provider must observe the care of a good contractor. Without specific agreements, a provider of services is therefore usually held to an obligation of best efforts. An SLA can go further by setting concrete results, such as a guaranteed availability. That makes the SLA legally important: it converts abstract promises into measurable commitments.
Like any contract, an SLA is interpreted not only on its wording but on what the parties could reasonably expect of each other in the circumstances. Clear definitions of terms such as “availability”, “incident” and “resolution” therefore reduce the room for argument. An SLA that is part of the provider’s general terms and conditions is also subject to the rules on general terms (Articles 6:231 and following BW).
Which performance metrics are common, and what do they mean legally?
Common metrics are availability (uptime) as a percentage, response time, resolution time and quality standards. Legally, each metric only has force if it is clearly defined and it is clear what happens if it is not met.
Take availability. An availability of 99.9 percent per month allows roughly 43 minutes of downtime; 99.5 percent allows about three and a half hours. The difference between these numbers matters, but so does the definition: is availability measured at the provider’s data centre or at the customer’s end, is planned maintenance excluded, and over which period is the percentage calculated?
If the provider does not meet an agreed service level, that is in principle a failure to perform (tekortkoming). The customer can then, depending on the contract and the law, claim performance, damages or dissolution of the contract (Articles 6:74 and 6:265 BW). In most cases, a notice of default setting a reasonable period for performance is required first (Article 6:82 BW), unless performance within the agreed time was essential. A well-drafted SLA regulates these consequences explicitly, so that the parties do not have to rely only on the general rules.
How does an SLA help manage risk?
An SLA helps both parties manage the risk of service disruptions. The customer knows what it can expect and what remedies it has; the provider knows exactly what is required and can limit its liability.
For regulated companies, it is sometimes a legal obligation. Financial institutions, for example, must include specific provisions in contracts with ICT service providers under the EU Digital Operational Resilience Act (DORA, Regulation 2022/2554), including precise descriptions of service levels for critical functions. Organisations covered by the NIS2 Directive must also manage risks in their supply chain, which affects the agreements they make with IT suppliers.
What are the key elements of an effective SLA?
An effective SLA contains precise metrics, clear consequences for missing them, a measurement and reporting method, and a procedure for review. Each element must be drafted so that it can be enforced.
How do you define measurable service levels?
Define each service level in objective terms: what is measured, how, by whom and over which period. Generic statements such as “high availability” or “prompt response” cannot be enforced.
A good SLA uses definitions and tables. For example: priority 1 incidents (complete outage of a critical system) must be responded to within 30 minutes and resolved within four hours, 24 hours a day; priority 3 incidents within one working day. Make clear whether times run only during business hours, what counts as “resolved” (a workaround or a permanent fix), and when the clock starts: when the customer reports the incident or when the provider detects it.
What consequences should follow if a service level is not met?
The SLA should state explicitly what happens when a service level is missed: service credits, an improvement plan, escalation and, for serious or repeated failures, a right to terminate. Without such provisions, you depend on the general rules of contract law.
Service credits are the most common remedy: a reduction of the fee, for example a percentage of the monthly fee for each percentage point below the agreed availability. Depending on their wording, service credits can qualify as a penalty clause (boetebeding) under Article 6:91 BW. That has an important consequence: unless otherwise agreed, a penalty replaces the statutory right to damages (Article 6:92(2) BW). If the customer wants to be able to claim damages above the service credits, the SLA must say so expressly. A court can reduce a penalty if fairness clearly requires it (Article 6:94 BW).
Balance is key. Service credits that are too low give the provider no incentive to improve; credits that are too high will be resisted by the provider or reflected in the price. Combine financial consequences with constructive measures, such as a root cause analysis and an improvement plan, and include a right to terminate if the service levels are missed repeatedly, for example in three consecutive months.
How do you keep the SLA up to date?
Include a periodic review, for example every quarter or every year, in which the parties discuss performance and adjust service levels where necessary. Record who may agree changes and how they are documented.
Make clear that changes to the SLA require written agreement signed by authorised representatives, so that there is no uncertainty about what applies.
How do you avoid common pitfalls in SLAs?
Most SLA problems come from vague definitions, unclear escalation, inadequate measurement and a lack of exit arrangements. You avoid them by checking the draft for these points before signing.
Which weaknesses should you look for?
Review the draft SLA for ambiguity in service definitions, metrics and legal consequences. Every point that can be read in two ways is a potential dispute.
Typical weaknesses are imprecise measurement, missing or unclear escalation procedures, an unclear relationship between the SLA and the main contract, and exclusions that are so broad that the service levels mean little in practice, for example a broad definition of “planned maintenance” without limits on duration or timing. Check also whether the limitation of liability in the main contract undermines the SLA: a very low cap on liability can make even a clear service level of little value.
The table below lists typical vulnerabilities and how to address them.
| Vulnerability | Example | Mitigation |
|---|---|---|
| Imprecise metrics | “High availability” without a percentage | Specific, quantifiable definitions |
| Inadequate dispute resolution | No agreed steps before going to court | Escalation steps with deadlines, choice of forum |
| Unclear escalation | Confusion about who to contact when | Step-by-step escalation procedure with names or roles |
| Insufficient monitoring | Only the provider measures, without access for the customer | Shared reporting, audit rights |
| Contractual ambiguity | Conflict between SLA and main contract | Order of precedence clause, legal review |
How do you monitor and verify performance?
Agree on how performance is measured and reported, and make sure the customer can verify the figures. A service level that only the provider can measure is difficult to enforce.
In practice, the provider usually measures with its own monitoring tools and reports monthly. The customer should have the right to receive these reports, to inspect the underlying data and, where necessary, to have an independent audit carried out. Agree what happens if the customer’s own measurements differ from the provider’s, for example that incidents reported by the customer are logged in a ticketing system and count for the calculation.
Discuss the monthly reports in regular service meetings and keep them: in a dispute, they are the main evidence of whether the service levels were met.
How do you prepare for disruption and exit?
Plan for the scenarios that matter: a serious outage, a data breach, the provider’s insolvency and the end of the contract. For each, the SLA or the main contract should state what the provider must do.
Include business continuity and disaster recovery obligations, such as recovery time and recovery point objectives, and the obligation to test them. Arrange data protection properly: if the provider processes personal data on your behalf, a data processing agreement is required under Article 28 of the General Data Protection Regulation (GDPR).
Arrange the exit in advance. The SLA should state that the provider must return your data in a usable, standard format, cooperate with the transition to a new provider and continue the service during a transition period. For cloud services, the EU Data Act (Regulation 2023/2854) has applied since 12 September 2025: customers may switch provider with a maximum notice period of two months, after which the provider must complete the transition within a transitional period of in principle 30 days. From 12 January 2027, providers may no longer charge switching fees. Where the provider could become insolvent, consider an escrow arrangement for critical software.
How do you negotiate a good SLA?
Prepare by defining your own needs, compare the provider’s offer with alternatives, and negotiate the service levels, the remedies and the exit together. The best SLA is precise where it matters and flexible where business needs will change.
How do you prepare for the negotiation?
Start with an internal assessment: which services are critical, what downtime can the business tolerate, and what does an outage cost per hour? The answers determine which service levels you really need and what they are worth to you.
Involve legal, IT, operations and finance in the preparation. Compare the provider’s standard SLA with market practice and with offers from other providers. Pay attention to the provider’s standard terms: they often contain broad exclusions and low liability caps that affect the value of the service levels. As a provider, assess realistically what you can deliver, and make sure your own subcontractors, such as data centre or cloud providers, offer at least the same service levels.
How do you negotiate the metrics themselves?
Negotiate metrics that are precise, measurable and linked to the business impact of the service. Build in a mechanism to adjust them as the service or your business changes.
Agree how service levels can be adjusted, for example after a change in the service, a migration or growth in usage. A change procedure with clear rules on price and timing prevents discussions later. For the customer, it is also important that service levels cannot be lowered unilaterally by the provider, for example through a change in its general terms.
How do you arrange governance and dispute resolution?
Agree on a governance structure with regular meetings, an escalation ladder and, if needed, an independent expert for technical disputes. Also agree on the applicable law and the competent court or arbitration.
A typical escalation ladder starts with the service managers, then moves to senior management and only then to mediation, expert determination or proceedings, with deadlines for each step. Make sure the escalation procedure does not prevent urgent measures, such as summary proceedings (kort geding) to force the provider to continue the service. For technical disputes about whether a service level was met, a binding opinion (bindend advies) from an independent IT expert can be quicker and cheaper than court proceedings.
In international SLAs, the choice of law and forum matters. Parties can choose Dutch law under the Rome I Regulation, and a choice of a Dutch court is in principle respected within the EU. Arbitration is an option when the other party is outside the EU.
What should you check as a customer?
- Check that availability is a defined percentage over a defined measurement period, not a “reasonable endeavours” promise.
- Check whether service credits are the exclusive remedy, or whether damages above the credits remain possible under Article 6:92 BW.
- Check that the liability cap in the main contract does not empty the service levels.
- Check your right to receive the monitoring data and to have an independent audit carried out.
- Check the exit: return of data in a standard format, transition assistance and, for cloud services, your switching rights under the Data Act.
What should you check as a provider?
- Check that each service level is achievable and backed by your own subcontractors’ SLAs.
- Define planned maintenance, force majeure and customer-caused incidents as exclusions.
- Cap the service credits, for example at a percentage of the monthly fee.
- Check whether your data-processing role requires a data processing agreement under Article 28 GDPR.
- Plan the Data Act switching obligations, including the 30-day transitional period.
What can we do for you with your SLA?
Our IT and contract lawyers draft and review SLAs and IT and outsourcing contracts for customers and providers.
- We review your draft SLA against the main contract and flag vague metrics, broad exclusions and low liability caps.
- We draft service levels as enforceable obligations of result, with measurement rules and exclusions.
- We calibrate the service credits and their relationship with damages under Articles 6:91 to 6:94 BW.
- We draft the exit, escrow and Data Act switching clauses.
- We send the notice of default and pursue performance, damages or dissolution when the provider falls short.
Summary
- An SLA is an ordinary contract under Dutch law; for services, Articles 7:400 and following BW apply alongside general contract law.
- Write service levels as obligations of result with defined percentages, measurement periods and exclusions; “reasonable endeavours” is much harder to enforce.
- Service credits can qualify as a penalty clause that replaces damages unless the SLA says otherwise (Article 6:92(2) BW).
- Agree who measures, how performance is reported and verified, and when repeated failures justify termination.
- Arrange the exit in advance; for cloud services, the EU Data Act gives customers switching rights since 12 September 2025.
Frequently asked questions
What is a Service Level Agreement (SLA)?
An SLA is a contract, or an annex to a contract, that sets out the service levels a provider must deliver, such as availability, response and resolution times, how these are measured and what happens if they are not met. Under Dutch law it is binding like any other contract.
Why are SLAs important for businesses in the Netherlands?
Without an SLA, a service provider is usually only held to the care of a good contractor (Article 7:401 of the Dutch Civil Code), which is hard to enforce. An SLA makes performance measurable, sets out the consequences of failures and arranges the exit, which gives both parties certainty and reduces disputes.
What key elements should be included in an effective SLA?
An effective SLA contains precise and measurable service levels, a measurement and reporting method, service credits and their relationship with damages, an escalation procedure, termination rights for repeated failures, a review procedure and exit arrangements, including the return of data.
