
Disaster Recovery Plan Template for Small Businesses
A disaster recovery plan answers four questions before the disaster happens: what can go wrong, what do we restore first, who does what, and how fast must it happen. The template below fits on a few pages and is written for a business without a dedicated IT team.
DR vs. backup: they are not the same thing
A backup is a copy of your data. Disaster recovery is the plan for using that copy under pressure: at 7am on a Tuesday, with the server dead, the phones ringing, and the one person who knows the passwords on leave. Most small businesses have backups; very few have written down what happens next. That gap is why identical incidents put one company back online by lunch and another out of business.
The enterprise templates you'll find elsewhere assume a DR coordinator, an IT department and a standby data centre. This one assumes you have none of those: just a small team, your backups, and a few honest decisions made in advance.
RTO and RPO in plain words
Two acronyms carry the whole subject, and both are just business questions:
RTO (Recovery Time Objective): "How long can we afford to be down?" If the accounting server takes two days to rebuild and you decided one day is the maximum tolerable, your plan fails before you start. Decide the number first, then check your restore process can meet it.
RPO (Recovery Point Objective): "How much work can we afford to lose?" If you back up nightly, everything since last night is gone when you restore. If losing a day of invoices is unacceptable, your RPO decision just changed your backup schedule.
Write both down per system, not company-wide. Your file archive can tolerate a day; your orders database probably can't.
Copy this template, formatted, to paste straight into Word or Google Docs.
The template
Copy the structure below into a document. It has five parts; keep the whole thing under five pages, because a plan nobody can skim in an emergency is a plan nobody follows. Replace the [bracketed] parts.
Disaster Recovery Plan: [Company Name]
Version: [1.0] · Owner: [name/role] · Last tested: [date] · A printed copy lives at [location outside the office].
Part 1: Scenarios we plan for
We plan for the effect, not the cause. Four situations cover almost everything:
- S1: One system dead (server or disk failure, corrupted database)
- S2: Ransomware / malicious action (systems up but data encrypted or destroyed; attacker may still be inside)
- S3: Premises unusable (fire, flood, theft, no access to the office)
- S4: Key person unavailable (the one who knows the passwords, during any of the above)
Part 2: Priority order of systems
Restore in this order. Decided in advance so nobody debates it during the incident.
| # | System | RTO (max downtime) | RPO (max data loss) | Restore runbook |
|---|---|---|---|---|
| 1 | [Orders / core database] | [4 h] | [1 h] | [link/appendix] |
| 2 | [Email] | [8 h] | [4 h] | [link/appendix] |
| 3 | [Accounting / payroll] | [1 day] | [1 day] | [link/appendix] |
| 4 | [Shared files] | [2 days] | [1 day] | [link/appendix] |
Part 3: Who does what (contact tree)
| Role | Responsibility |
|---|---|
| Incident lead: [name, phone] | Declares the incident, makes the restore-order calls, is the single decision-maker. Deputy: [name, phone]. |
| Technical restore: [name/IT provider, phone, after-hours number] | Executes the runbooks. Deputy: [name]. |
| Communications: [name, phone] | Tells staff and customers (see Part 5); nobody else communicates externally. |
| Credentials | Emergency access to passwords/keys is via [password-manager emergency access / sealed envelope / second keyholder]. This is the answer to scenario S4. |
Part 4: Restore runbooks (one per system, one page each)
Each runbook, written so the deputy can follow it: where the backups are (primary and offsite); how to get the restore target (spare machine, cloud VM, replacement hardware source); the restore steps; how to check the restore is good (what to open, what number to compare); measured restore duration from the last test: [hh:mm].
Ransomware addition (S2): do not restore over the top: isolate the network first, restore to clean hardware/VMs from a copy that predates the infection, and change credentials before reconnecting. The immutable offsite copy is the one you can trust.
Part 5: Communication
| Audience | Message |
|---|---|
| Staff | Via [WhatsApp group / phone tree]: what's down, what to use meanwhile, when the next update comes. |
| Customers | A short honest holding message, prepared in advance: ["We are experiencing a technical issue affecting X. Your data is safe. We expect normal service by Y and will update you at Z."] |
| Regulatory | If personal data may be lost or breached, assess notification duties (e.g. under PDPA's breach-notification requirements); owner: [name]. |
The annual test: two hours, no server harmed
A plan that has never been tested is a guess. Once a year, run a tabletop exercise: book two hours, pick a scenario ("the office flooded last night"), and walk the room through the plan on paper: who declares, who calls whom, which runbook opens first, where the passwords come from. Note every point where someone says "actually, that's changed". Those notes are the real output. Separately, restore tests of individual systems (Part 4's measured durations) should happen far more often than annually; that part is automatable.
Keeping it current
A DR plan rots quietly: servers get replaced, people leave, backup destinations move. Two habits keep it alive: review the plan whenever the backup register changes, and keep the restore-duration numbers in Part 4 from real, recent tests, not estimates. This is where Backup Governor earns its place in a DR plan: scheduled real restore tests with measured durations turn your RTO from a hope into a number you've hit before, and continuous verification tells you before the disaster that the backup you're counting on is restorable. See How to Prove Your Backups Actually Work for the evidence side.
Turn your RTO into a measured number
Backup Governor runs scheduled restore tests and records how long each one took, so Part 4 of your plan fills itself in.
Get started free