Most small businesses have a backup. Very few know what to do, and in what order, on the morning after an attack.

You only find out when it happens. The screens show a ransom note, the managing director's mailbox is locked, and nobody can find the number for the IT partner, because it lives in the address book on the encrypted server.

An IT disaster recovery plan does not decide whether an attack succeeds. It decides how quickly you are working again afterwards. Its core answers fit on one page, and you can put it together in two afternoons.

What an IT disaster recovery plan is, and what it is not

An IT disaster recovery plan sets out who does what after a major IT outage, in which order systems come back, and how your business keeps going in the meantime. The trigger might be a cyber attack, but just as well a fire, a failed server or a supplier that goes down.

Three terms tend to get mixed up:

  • A backup is a copy of your data. It is the precondition for recovery, not recovery itself.
  • Disaster recovery is the technical route from that copy back to normal operations.
  • Business continuity management is the wider frame: how the business stays able to act whether it is IT, a supplier or a building that fails. The international standard for it is ISO 22301.

Prevention is not the subject here. The gaps attackers exploit most often in small businesses are covered in cyber security for SMEs. This article deals with the question that comes after: what happens when protection fails once?

Why this concerns small businesses too

In the second half of 2025, 57 ransomware incidents were reported to the National Cyber Security Centre (NCSC), the federal office for cyber security. That figure comes from its semi-annual report 2025/2, published on 30 March 2026. It counts reported cases only; nobody knows the true number.

On top of that, an obligation has applied since 1 April 2025: operators of critical infrastructure must report cyber attacks to the NCSC within 24 hours. Most small businesses are not directly covered. But if you sit in such a supply chain, say as a supplier to a municipality or an energy utility, sooner or later a customer will ask whether you have a recovery plan.

Two numbers first: how long and how much

Before you write a single line of the plan, answer two questions for every important process:

  • How long may it be down? That is the recovery time objective (RTO).
  • How much work may be lost? That is the recovery point objective (RPO). If you back up daily at 02:00, the worst case is a full working day lost.

The answers are business decisions, not technical ones. For a trades business with twelve staff, a first pass might look like this:

ProcessMay be down forMay lose
Phone and email4 hoursnothing
Job scheduling and appointments1 day1 hour
Quotes and invoices2 days1 day
Payroll1 week, except just before payday1 day
Website2 days1 week

The values are an example, not a benchmark. What matters is that they are set before anything happens. Only then can you check whether your backup can meet them at all. A daily backup does not fit "nothing may be lost".

What goes into the plan

A usable plan is short. It answers seven questions:

  1. Who decides? One person leads the response and a second stands in. Both are named.
  2. Who do you call? IT partner, insurer, bank, cantonal police, your most important customers and suppliers, with phone numbers.
  3. In what order do systems come back? The table above, sorted by priority.
  4. How do you communicate when email and Teams are down? A phone tree or a group chat on personal phones, agreed in advance.
  5. Where are the emergency credentials? At least one administrator account that is not tied to everyday work, with two-factor authentication, kept sealed.
  6. How does the business run without IT? Paper forms for orders, a list of open appointments, a way to tell customers what is going on.
  7. Who do you have to notify? More on that below.

The plan also exists on paper. A recovery plan stored only on the server gets encrypted along with everything else. One printed copy with the person in charge and one at home is enough.

The first hours after an attack

On its ransomware page, the NCSC sets out the steps for affected organisations. For a small business they fall into this order:

  1. Stay calm and alert the person in charge. Rushing destroys evidence and costs time.
  2. Disconnect affected devices from the network: pull the network cable and switch off Wi-Fi. Do not restart them.
  3. Bring in specialists if you lack the expertise in-house.
  4. Secure evidence before anything is repaired. Specialists image memory and disks before anyone restarts a machine. After that, an investigation is barely possible.
  5. Check backups before restoring. The NCSC names finding the infection path as a key goal. Only once you know how the attacker got in, and since when, do you know which backup is clean.
  6. Rebuild rather than repair. The NCSC recommends reinstalling infected systems and only then restoring data from backup.
  7. Notify and report. A criminal complaint goes to the cantonal police where the business is registered.

On paying the ransom, the advice is clear: the NCSC advises against it. There is no guarantee the data will not be published anyway. Anyone still considering payment should discuss it with the cantonal police first.

Who you have to notify, and when

  • Personal data affected: if the data security breach is likely to result in a high risk for the people concerned, you must report it to the Federal Data Protection and Information Commissioner (FDPIC) as soon as possible (Art. 24 FADP). What else the act requires is covered in AI and data protection in Switzerland.
  • Critical infrastructure: report to the NCSC within 24 hours.
  • Insurance: if you have cyber insurance, your policy states how and by when a claim must be reported. That deadline belongs in the plan.

This overview is not legal advice. It only shows which questions will be on the table on day one.

Backups that survive the real thing

If backups can be reached from the same account as everyday work, an attacker with administrator rights can encrypt or delete them too. Then there is nothing left to restore.

The NCSC recommends the 3-2-1 rule: three copies of your data, on two different storage media, one of them off site. Ransomware adds a requirement: at least one copy must sit where an infected system cannot delete it, offline or in locked storage.

Three questions to put to your IT partner:

  • Which account can delete the backups, and is it separate from everyday accounts?
  • How far back do the backups go? An attacker may have been in the network for some time before encrypting anything. The latest backup may then not be clean.
  • Are email, OneDrive and SharePoint in Microsoft 365 backed up too, and outside the account itself?

Where the data should live in the first place, in your own server room or in the cloud, is covered in on-premise vs cloud.

A plan never rehearsed is an assumption

A backup that has never been restored is a promise without proof. The NCSC advises testing restores at intervals. A restore test is only worth something once it is recorded:

  • the date
  • what was restored: a file, a mailbox, a whole server
  • how long it took
  • whether the result was complete

Duration is the number that matters most. It shows whether the recovery time in your table is realistic.

Add a tabletop exercise once a year: ninety minutes, the person in charge, the IT partner and a scenario such as "Monday, 07:30, every drive is encrypted". You walk through the plan step by step. Whatever turns out to be missing is the most valuable finding of the year.

What ALPENIQ does here

ALPENIQ IT sets up the technical side: automatic backup of every business-critical system, a copy in a Swiss data centre that is locked against deletion, backup for Exchange, OneDrive, SharePoint and Teams, and agreed intervals and recovery targets. Restore tests with a written record are included, as is a recovery plan for the loss of entire systems. The scope is under Backup & cloud, protection against attacks under IT security.

One limit belongs here too: which processes come back first, and how long your business can manage without them, is not something anyone outside can decide for you. That table is for management to fill in.

Conclusion

An IT disaster recovery plan does not need to be thick. It needs two numbers per process, seven answers on one page, backups an attacker cannot reach, and a test that shows how long a restore really takes. Set that before an incident, and an incident costs you hours instead of weeks.

Further reading

Frequently asked questions (FAQ)

What goes into an IT disaster recovery plan?

An IT disaster recovery plan names the person who decides in an emergency and their deputy, lists every important contact with phone numbers, and sets the order in which systems come back. It also covers a way to communicate without email, the emergency credentials and the reporting duties. It exists on paper as well, because otherwise an attack encrypts it too.

What is the 3-2-1 backup rule?

The 3-2-1 rule calls for three copies of your data on two different storage media, one of them off site. The Swiss NCSC recommends it explicitly. Against ransomware you also need at least one copy that an infected system cannot delete, for instance offline or in locked storage.

What is business continuity management?

Business continuity management makes sure a company can keep acting through any major disruption, whether IT, a supplier or a building fails. The IT disaster recovery plan is one part of it. Organisations that want to set it up formally can use the international standard ISO 22301.

What should you do after a ransomware attack?

Disconnect affected devices from the network, bring in the person in charge and a specialist, secure evidence before anything is restarted, and rebuild systems rather than repair them. Then report: to the FDPIC if personal data faces a high risk, to the NCSC if you operate critical infrastructure, and file a complaint with the cantonal police. The NCSC advises against paying the ransom.

How often should a small business test its restores?

The NCSC sets no fixed interval; it advises testing restores at intervals. A practical rhythm for a small business is a small test each quarter, a single mailbox for instance, and a larger one each year alongside a tabletop exercise. What counts is the record: date, what was restored, how long it took and the result.