top of page

DORA is almost two years old: what has the Luxembourg market of service providers learned?

12 minutes ago
4 min read

The EU’s Digital Operational Resilience Act (DORA) has been in application since 17 January 2025.
For Luxembourg financial entities and service providers, the first implementation cycle has shown that effective compliance depends less on the volume of documentation than on whether governance, risk decisions and operational evidence form a coherent and defensible system. Implementation of requirements has been considered demanding, but Luxembourg service providers have coped with this challenge well.
 
DORA was designed to strengthen the ability of EU financial entities to manage ICT risk, respond to incidents, test operational resilience and oversee ICT third-party dependencies. Its central achievement is to place digital operational resilience firmly within the accountability of the management body rather than treating it as a purely technical discipline. 
 
Implementation was never likely to be uniform. Article 4 requires proportionality by reference to the entity’s size, overall risk profile, and the nature, scale and complexity of its services, activities and operations. The practical lesson, however, is that proportionality affects the depth and sophistication of implementation, not whether the relevant risk domains are addressed.

Proportionate, not insufficient


Smaller or less complex entities may reasonably use less automated monitoring, simpler governance arrangements or less elaborate testing methodologies. Those choices must nevertheless be based on an explicit assessment of risk and criticality. Resource constraints, headcount or budget, by themselves, do not constitute a defensible proportionality methodology. 
 
A stronger approach is to define proportionality criteria in advance and apply them consistently across ICT risk management, incident management, resilience testing and ICT third-party oversight. Relevant factors include the criticality and substitutability of services, ICT architecture, outsourcing model, systemic relevance and the entity’s operational dependencies. Decisions should be approved through the appropriate governance process, reviewed periodically and revisited when the business or ICT environment changes materially. 
Consistency is essential. A service, system or dependency should not be treated as low risk in one process but critical in another without a documented explanation. Compliance and internal audit should therefore challenge whether classifications, testing frequencies and oversight intensity remain coherent across the DORA framework. 
L3A recommends avoiding generic DORA templates where possible. Service providers should first look at their operating models and risk profiles, and define their proportionality criteria before selecting which controls to apply.
Any such frameworks should include a business impact analysis, ICT risk register and asset inventories. Incident data, testing results, audit findings and a register of information can be added over time. Each provider should be able to explain the chain from business function, via technology dependency and risk scenarios, to test evidence and remediation.

Effective governance, from documentation to decision-making 


So has DORA lifted ICT risk from a purely IT issue to a genuine management responsibility? It certainly should have, but implementation can be patchy.
DORA requires the management body to define, approve, oversee and remain responsible for the implementation of the ICT risk management framework. This responsibility is not discharged by receiving technical dashboards or status updates. Reporting must enable management to understand the entity’s current risk posture and make identifiable decisions. 
The L3A Working Group recommends regular, decision-oriented management reporting. Depending on the entity’s risk profile, this may include the external threat landscape, ICT incidents and trends, top ICT risks, testing outcomes, overdue findings, third-party dependencies, concentration risk and progress against remediation plans. Explicit decision points should be highlighted, including policy approvals, acceptance of residual risk, prioritisation of remediation, investment decisions and escalation of material weaknesses. 
The periodic report on the review of the ICT risk management framework serves a different purpose. It should provide a consolidated assessment of the framework’s effectiveness, proportionality and maturity, drawing on sources such as ICT risk assessments, incident data, audit findings, resilience testing, the Register of Information and management reporting. Under the applicable delegated rules, it must follow the prescribed structure and be available to the competent authority on request. 

For this, the L3A guidance on DORA https://www.l3a.lu/post/dora-practical-industry-driven-implementation-guidanc recommends at least quarterly reporting, with a view on current and future external threats, as well as full incident and response events and trends.
Most importantly, reports should identify explicit decision points, such as policy approval, residual-risk acceptance, remediation priorities, investment decisions or escalation. The quality test is whether this information enables decision-making, not simply whether it is comprehensive.

Fragmentation and concentration – the operational evidence chain 


So far, evidence suggests that there are recurring weaknesses due to both fragmentation and concentration. Internally, ICT components and frameworks may be disconnected, and external visibility of third-party providers can be low, leading to fragmented ICT audits, processes and weaker overall insight and control.
Third-party oversight remains one of the most difficult areas in practice. Financial entities depend on extensive vendor ecosystems, cloud providers and multi-layer subcontracting chains, while their ability to negotiate bespoke audit rights or enforce requirements throughout the chain may be limited. DORA nevertheless requires accurate provider information, appropriate contractual protections, ongoing monitoring and viable exit strategies for ICT services supporting critical or important functions. 
Alternative assurance mechanisms, including pooled audits, independent assurance reports and recognised certifications, can support oversight where direct audits are impractical, but they must be assessed for scope, relevance and sufficiency. They do not remove the financial entity’s responsibility to understand the service, its dependencies and the residual risk. 
Concentration risk must also be considered at portfolio level. The relevant question is not only whether an individual contract contains an exit clause, but whether substitution, data portability, recovery capability and continuity arrangements are genuinely feasible across the entity’s ICT ecosystem. Where diversification is not practicable, the resulting risk should be explicitly recorded, monitored and mitigated. 
Should the CSSF visit tomorrow, providing a large stack of documents is unlikely to satisfy. The strongest proof will be traceable evidence that the firm can and does detect weaknesses, makes the right decisions, remediates them and verifies that critical services can continue or recover.

What good implementation now looks like


After almost two years of application, the principal lesson is that DORA is not a one-off documentation exercise. It is a governance and operating model that requires continuous maintenance, testing, learning and improvement. A large set of policies will not, by itself, demonstrate resilience. 
The strongest supervisory position is a coherent evidence base showing that the entity understands its critical functions and dependencies, detects and escalates incidents, tests realistic disruption scenarios, takes accountable decisions, remediates weaknesses and verifies that critical services can continue or recover. That is the point at which digital operational resilience becomes more than compliance and starts to influence how the business is managed. 

Article written by VitalBriefing

Published in November 2026

Comments


bottom of page