GUIDE · DORA

ICT risk management framework under DORA: structure and content

After this guide you will know what content the ICT risk management framework required by DORA must include, how to organise the document so that it can be verified, and the order in which to write its parts. The framework is the basis on which the regulation's other obligations rest, so it is worth setting it up well from the start.

Updated october 2026
8 min read
Altovar Technical Editorial Team

1. Who must have it, and why

The ICT risk management framework is required of financial entities within the scope of DORA. The answer depends on the type of authorisation and the activity carried out. The regulation applies a principle of proportionality, so a small entity may have a leaner document. Do not, however, assume an exemption simply because you are small: verify with the compliance function or a legal adviser.

The document serves two purposes. The first is internal: to give the governing body a clear view of ICT risks and responsibilities. The second is external: to show supervisors a systematic rather than episodic approach to management. A framework written only to answer an inspection quickly becomes obsolete, because nobody uses it to decide. A living document serves decisions, not just inspectors.

2. What the regulation requires in practice

DORA requires the framework to cover ICT risk strategy, policies and procedures, internal control, incident management, continuity and recovery. It does not impose a single format, but the content must be recognisable and linked together. Each component explains how decisions are made about what to protect, how the protection is verified and who decides. Clarify also who updates it.

The management body approves the framework and states the risk tolerance: how much service disruption is acceptable and for how long. Without this explicit decision, technical choices have no common criterion. The table shows the main components and the evidence you should be able to produce for each. Read each row as a question the document must answer.

ComponentQuestion it answersEvidence to keep
Strategy and governanceWhich risks do we accept and who is accountable for themApproval minutes and risk tolerance
Inventory of assets and servicesWhich systems, data and suppliers support our processesInventory with date of last update
Protection and detectionWhich measures protect our systems and how an anomaly is spottedConfigurations, monitoring logs and outcome of access reviews
Response and recoveryWhat happens when something goes wrong and how we return to operationApproved procedures and recovery test reports
Review and learningHow the framework is updated after incidents and changesVersion history and corrective action register

3. Recommended structure of the document

A clear structure helps both those who write and those who review. One possible organisation follows the lifecycle of risk: objectives and scope, governance, system inventory, measures and monitoring, response and recovery, review. Each chapter names its owner and the date of its last review. Use the same layout for every chapter, so that readers always find the same references and the same questions answered.

Avoid documents of hundreds of pages that nobody reads. A main document of a few dozen pages, with operational annexes such as procedures, the risk matrix and the list of controls, is better. Readers must be able to find the answer to their question within minutes, otherwise the document will not be used. Include an index with useful internal references.

  1. Objectives, scope and risk tolerance approved by the governing body.
  2. Roles and responsibilities, with a matrix of who decides, who executes and who verifies.
  3. Inventory of ICT systems, data and services, linked to business processes.
  4. Protection measures, with the rules for access, configuration and change management.
  5. Detection of anomalies and monitoring of critical systems.
  6. Incident response management and recovery plans.
  7. Backup policies and periodic verification of restores.
  8. Periodic review and lessons learned from events.

4. The evidence to keep

A framework is worth as much as the evidence that supports it. For each chapter, keep at least one verifiable item: the approval minutes, the inventory with the date of its last update, the report of a recovery test, the outcome of an access rights review. Without this evidence, the document describes intentions rather than practice. There need not be a great deal of evidence, but it must be retrievable within minutes.

5. Order of work

Not everything should be written at once. The starting point is the inventory, because without the list of systems, services and suppliers every chapter remains abstract. Immediately after come governance and risk tolerance, which set the criteria for all subsequent choices. Anyone who skips these two steps ends up rewriting the measures twice. An inventory produced after the measures almost always leads to the wrong measures.

6. Common pitfalls

The most frequent errors are substantive rather than formal. Before presenting the framework, check that none of these problems is present. Many arise from rushing to close the document for a deadline, and they become obvious on a careful reading. If an outsider cannot follow it, neither will an inspector. Give this stage an extra review cycle.

Approval minutes with no discussion do not demonstrate governance. Record in the minutes the questions raised, the tolerances discussed and the decisions taken, even in summary form. You do not need a long text: concrete traces of what was decided and why are what make the process credible. The reader should be able to understand why the decision was taken.

  • Writing a generic document, copied from a template, that does not describe the organisation's actual systems.
  • Declaring protection measures that nobody applies or that cannot be verified with evidence.
  • Treating the framework as separate from the supplier register and the continuity plan, when they are parts of the same system.
  • Not stating who updates the document after approval, and how often.
  • Approving the framework without a real discussion at board level, simply by signing it.

7. Points to clarify with your adviser or the authority

Some points depend on your specific situation and must be confirmed with a legal adviser or the compliance function. These include whether your entity falls within scope and with which obligations, how to apply the principle of proportionality in practice, what the expectations of the supervisory authority for your sector are, and how the framework links to the other documents the regulation requires.

Also ask the suppliers that support your systems which evidence they can provide, because part of the framework depends on their work. Put the answers in writing and keep them with the rest of the evidence. A supplier that cannot indicate its own evidence is a signal to record in the framework, because it points to a risk to manage. Also ask for a response date.

8. How to get started

This week, prepare the minimum inventory: a list of systems, ICT services and the suppliers that manage them, with the internal owner of each. It need not be complete, but it should be precise enough to become the basis of the assets chapter. Ask each owner to confirm their line by an agreed date. If someone does not respond, record it in the document anyway as an open gap.

Then arrange a meeting with the governing body, or with those who represent it, to discuss risk tolerance. Bring a written proposal with a few scenarios of disruption to the most critical services and ask for an explicit decision. That decision is the first real content of the framework, and everything else follows from it. Write up the minutes with the decision and the date straight away.

Frequently Asked Questions

Can the ICT risk management framework be built from existing documents?

Yes, and this is often the most efficient route. Many organisations already have security policies, continuity plans and IT procedures in writing. The work consists of checking them against the required content, filling gaps and linking them into a single framework with clear governance. Reuse is fine, provided the result is consistent.

Does the framework need to be reviewed every year?

It should be reviewed periodically, at a frequency defined in the document itself. It is common to provide for at least an annual review and extraordinary updates after significant incidents or major changes to systems and suppliers. The frequency should be chosen according to the complexity of the organisation and confirmed with the compliance function.

Is a dedicated chapter on external suppliers needed?

Yes. Risk linked to ICT suppliers is part of risk management and should be connected to the contract register. The framework must explain how you assess a supplier before choosing it, how you monitor its performance over time and how you manage the end of the relationship, including a move to another supplier.

What is the difference between a policy and a procedure?

A policy states what must be done and why, and assigns responsibilities. A procedure describes the concrete steps to do it. For example, an access policy establishes that privileged access must be approved and reviewed; the procedure explains how to request it and who approves it. Keeping the two levels separate makes the document easier to update.

Must the framework be published?

Normally it is an internal, confidential document. It must, however, be made available to the supervisory authority when requested, and it is useful for it to be known to those who must apply it. Ask your adviser which parts should be circulated internally and which should remain confidential, because circulation may involve security details that should not be exposed.

About this content

General informational guide. Regulations and technical requirements change over time: always check the official sources and assess your specific case with a qualified technical or legal advisor.

GET IN TOUCH

Talk to our team.

We'll show how Altovar integrates into your infrastructure. No commitment, no sales pitch. Just a technical conversation.

BOOK A DEMO →