1. Why the first hours determine the cost of the attack
Ransomware is software that encrypts data, that is, makes it unreadable, and demands a ransom to restore access. In the first hours the attack is often still limited to a few machines, while the code that spreads it works quietly across the network. Every minute in which the malware can move between systems adds encrypted data, stopped services and recovery work. For this reason the first phase matters more than any decision made later.
The difficult part is that the first hours coincide with confusion. Whoever notices the anomaly may think of a fault, a faulty update or a network problem. A playbook written in advance shortens this gap, because it tells whoever spots the anomaly what to do and whom to call, without wasting time debating whether the attack is real.
2. What to prepare before it happens
The first hours go well only if certain conditions are already in place. You need an up-to-date list of systems, with the owner of each and an indication of which are critical to the business. You need reachable contacts for the technical lead, management and the IT supplier, including out of hours, and a communication channel that does not run through company email, which may be compromised. Without these elements, every minute is spent looking for a number or a person.
You also need to know in advance which copies exist, where they are and how separated they are from the network. A backup permanently connected to the systems can be encrypted together with the original data. The 3-2-1 rule is a useful starting point: three copies, on two different media, one of them off-site. It is also worth checking that at least one copy is immutable, meaning that nobody, not even an administrator, can delete or alter it for a defined period.
- Inventory of systems with owner and criticality
- Up-to-date emergency contacts, including out of hours
- An alternative communication channel to company email
- Separate, immutable and tested backup copies
- Written procedure for isolating systems
- List of suppliers with access and how they intervene
3. Recognising the signs of ransomware
The most common signs are files that change extension or become unreadable, ransom demands appearing on screen or as text files, sudden slowdowns on several machines at once and unusual activity on administrative accounts. The sudden lock-up of many workstations at the same moment should also be taken seriously. None of these signs on its own proves ransomware, but together they justify activating the playbook.
The first step is not to seek certainty before acting. Anyone who detects an anomaly must report it straight away to the technical lead, with the time, the system and the symptom observed, without switching off or restarting the machine out of curiosity. Even a false alarm costs little compared with an attack that spreads while people debate whether it is real. The report should be simple, even spoken or a short message, because at that moment nobody has time to fill in forms.
4. The first-hour actions, in order
The following sequence is a starting point. It must be adapted to your systems and to the decisions your environment requires, and it should be tested before it is needed. Each step has the same purpose: stopping the spread, preserving what is needed to understand what happened and preparing recovery without improvising. The order matters: containment comes before in-depth analysis, which in turn comes before any attempt at recovery.
Switching off a machine is a decision to be taken with the technical lead. Usually it is best to disconnect it from the network and leave it powered on, because memory may hold information useful for analysis. In other cases the spread is so rapid that stopping the system is necessary, for example when other systems are already involved and there is no way to isolate them in time. The playbook must say who decides and on the basis of which criteria, without leaving the choice to the person standing in front of the machine.
- Alert the technical lead and management, and open the action log with the time and name.
- Isolate the affected systems from the network, disconnecting them without deleting anything.
- Block suspicious administrative accounts and revoke active sessions where possible.
- Also isolate shared systems, such as file servers and connected backup devices, if the malware can reach them.
- Photograph the screen showing the ransom message and note the files and extensions involved.
- Check with the IT supplier that the backups have not been encrypted or deleted before touching them.
- Call the legal adviser and the person who assesses notifications, without yet deciding on payment.
5. Mistakes that make the incident worse
The most serious damage in the first hours almost always comes from well-meant actions. Restarting systems to make them work again erases memory useful for analysis and can make traces of hidden processes disappear. Trying to decrypt files with tools found online can make the situation worse, and deleting the ransom message removes valuable information needed to understand the attack.
Another mistake is starting recovery before understanding how the attacker got in. If the entry point is still open, the malware returns and the work done is lost. Even a hasty restore from an unverified copy can bring the problem back together with the data. Finally, avoid improvised communications with whoever sent the ransom demand, because they may confirm your business is a target of interest and complicate every later decision.
6. How to check that containment holds
After the first actions, check that the spread has really stopped. It is not enough that the visible symptoms have disappeared: the malware may be waiting on other systems or on accounts that have not yet shown signs. The check must be systematic, with an owner recording what was verified, when and with what result. A spot check is not sufficient when there are shared systems or accounts with broad privileges.
The check is complete when the response group can explain, with evidence, how access was blocked and which systems were verified. Until then, containment is a hypothesis, not a fact. Keep the logs exported before they are overwritten by automatic rotation, and note who exported them and when. These logs serve both to understand the attack and to demonstrate, later, the actions taken.
| Check | What it indicates | Where to look |
|---|---|---|
| New machines with encrypted files | The spread has not stopped | Monitoring and user reports |
| Unusual administrative access | An account has been used by others | Access logs for privileged accounts |
| Unknown processes running | The malware may still be active | Control tools on the affected systems |
| Abnormal connections to external destinations | Possible data theft or remote command | Firewall and outbound traffic logs |
7. Recovery: what to check before using backups
Restoring from backups is the goal, but not every copy is useful. Verify that copies predate the infection, that they have not been encrypted and that restored systems do not carry the same entry point with them. A restore done too quickly can reinstall the problem together with the data. It is best to start with a test in an isolated environment before bringing data back into production.
Rebuild critical systems from clean images where possible, and restore data only after checks. Write the recovery order based on what is needed to keep the business going, not on what is easiest to bring back. The time to return to normal depends heavily on how well the recovery was prepared, and it varies with the scope of the attack.
8. Notifications and decisions that cannot wait
A ransomware attack can affect personal data. In that case GDPR, the European regulation on data protection, requires notification to the supervisory authority within 72 hours of becoming aware of the breach, when a risk to individuals is likely. If your business falls within the scope of the NIS2 directive on the security of network and information systems, transposed in Italy by Legislative Decree 138/2024, the first report must be made within 24 hours and the incident notification within 72 hours.
Paying the ransom is a decision the playbook must not take for granted. It does not guarantee data recovery, may not be the only option and requires legal and technical assessment. Decide in advance who assesses it and which information is needed to do so. Ask a legal adviser how these obligations apply to your situation, and record the decision with its reasons, because it will be needed to account for the choices made.
9. How to start
This week, write on a sheet the first steps you would take if the file server were found encrypted tomorrow. It does not have to be perfect: its purpose is to reveal which contacts, accesses and copies do not yet exist. Also note who, in your organisation, might block an action by mistake, because that is a detail the playbook must anticipate. The questions that emerge are the best material for the next piece of work.
Then check two things with your IT supplier: who responds out of hours, and whether there is at least one backup copy separate from the network and not modifiable. If either answer is uncertain, that is the first job to do. Also ask in writing which recovery times it can commit to for critical systems, because without a written answer the playbook remains a hypothesis. The rest of the playbook is built from there.
Frequently Asked Questions
What must I do first if I discover ransomware?
The first thing is to alert the technical lead and note the time, system and symptom. Straight after, isolate the affected systems from the network, without restarting them or deleting anything. Do not try to decrypt files yourself and do not reply to whoever sent the message. The playbook should state who decides the next steps, because confusion is high at that moment.
Should I switch off infected computers?
Not always. Isolating from the network is usually the first move, while switching off can lose information useful for later analysis. In some cases, however, the spread is so rapid that stopping the machine becomes necessary. Decide in advance with the technical lead and the supplier which approach to follow, and write it into the playbook.
Will backups always save me?
Not necessarily. A useful backup is recent, separate from the network, not modifiable and tested. If it is connected to the affected systems it can be encrypted together with the originals. The restore test is the only way to know whether a copy really works, and it should be carried out regularly, not only after an incident.
Should I pay the ransom?
The playbook should not give a fixed answer. Paying does not guarantee the return of data or the absence of further attacks, and it may have legal implications. The decision must be assessed with legal and technical advisers, taking into account notification obligations and the guidance of the competent authorities. Whoever decides should already be identified before an incident happens.
When must I notify the data protection authority?
It depends on the scope. For personal data, GDPR requires notification to the supervisory authority within 72 hours of becoming aware of the breach, when a risk to individuals is likely. If you fall under NIS2, the first report must be made within 24 hours. Not every case is the same. Check with your legal adviser which obligations apply to your business.
How long does it take to get back to normal?
There is no duration that applies to everyone. It depends on the extent of the attack, the quality of the backups, the ability to rebuild systems from clean images and the checks required. A prepared and tested plan reduces both the time and the uncertainty. Ask your IT supplier which recovery times it can commit to in writing for critical systems.