Disaster Recovery and Business Continuity Plan: Steps
Most small and mid-size companies discover their recovery gaps during the outage, not before it. This guide walks through building a disaster recovery and business continuity plan you can actually execute: setting RTO and RPO targets, pricing downtime, structuring backups, and testing the plan on a
It is 6:40 p.m. on a Friday when the warehouse management system stops responding. Orders stop printing. The pick team stands around. Someone calls the IT lead, who calls the hosting provider, who opens a ticket and promises an update in four hours.
Nobody in that chain knows two things that matter enormously: how long the company can tolerate being down, and how much data it is willing to lose. Those two numbers are the spine of a disaster recovery and business continuity plan, and most small and mid-size companies are forced to invent them during the outage instead of deciding them in advance.
This guide walks through building that plan as an operational document — targets, math, checklists, and a testing cadence — rather than a binder that lives in a shared drive and gets opened once a year during an audit.
What a Disaster Recovery and Business Continuity Plan Actually Covers
The two halves of the phrase are not synonyms, and conflating them is where plans quietly go wrong. Teams buy backup software, check the box, and assume the business is covered.
Disaster recovery versus business continuity
Disaster recovery is the technical half: restoring systems, data, and infrastructure after a failure. Backups, replicas, failover regions, DNS cutover, runbooks.
Business continuity is the operational half: keeping the company shipping, invoicing, and answering customers while the technical half is still in progress. Paper pick lists. A phone tree. A rule about which orders get hand-processed first.
A distribution company with flawless backups and no continuity plan still loses a day of shipments, because nobody decided who authorizes manual picking or how those transactions get reconciled later. FEMA's Ready.gov business continuity guidance is a reasonable free starting template if you have nothing written down at all.
RTO and RPO, in plain numbers
Two acronyms carry most of the weight. Define them per system, not for the company as a whole.
- RTO (Recovery Time Objective) — how long a system can be down before the damage becomes unacceptable. "Four hours." "Fifteen minutes." "Two days, honestly."
- RPO (Recovery Point Objective) — how much data you can afford to lose, measured in time. A nightly backup means an RPO of up to 24 hours. If an order entered at 4 p.m. must survive a 5 p.m. failure, your RPO is minutes, and nightly backups will not get you there.
Write these as a table with one row per system. Order entry, accounting, the customer portal, the internal wiki, and the marketing site do not deserve the same targets or the same spend. Treating them equally is how budgets get wasted protecting things nobody would miss until Monday.
The US National Institute of Standards and Technology publishes SP 800-34, its contingency planning guide for federal information systems, which is heavier than most mid-size companies need but useful as a structural reference for the business impact analysis section.
Build the Plan in Five Steps
Here is a sequence that works for a team of a few dozen people and does not require a dedicated continuity manager.
- Inventory what runs the business. List every system an employee touches in a normal week, including the ones IT did not buy. The spreadsheet on someone's desktop that calculates commissions counts.
- Assign RTO and RPO per system. Do this with the department head who owns the process, not alone in IT. Operations will tell you which outage actually stops revenue.
- Price the downtime. See below. This is the step that unlocks budget.
- Design the recovery path for each tier. Tier 1 might get a warm standby in a second region. Tier 3 might get a nightly backup and a documented four-hour rebuild. Both are valid answers.
- Write the runbook and name the humans. Specific names, specific phone numbers, specific decision authority. "The IT team" is not a name.
Price the downtime before you price the fix
Executives approve recovery spending when it is compared against a loss, not against a feature. Use a simple formula:
Hourly downtime cost = (gross revenue per operating hour × share of revenue the system touches) + (fully loaded hourly labor cost × employees blocked) + recovery labor + contractual penalties
Run it with your own numbers. If a company like Acme Logistics does $18,000 of gross revenue in an operating hour, the order system touches 80% of it, and 22 warehouse staff at $31 fully loaded sit idle, that is roughly $14,400 plus $682 per hour, before overtime to catch up and before any late-delivery penalties. Four hours of downtime crosses $60,000. Suddenly a standby environment that costs a few hundred dollars a month is not a hard conversation.
Be honest about the soft costs you cannot quantify, and label them as such instead of inventing a figure. Customers who place the next order elsewhere are real, but a made-up percentage will get your whole analysis dismissed.
The 3-2-1 rule, and why backups are not recovery
The durable guidance here is still the 3-2-1 rule: three copies of the data, on two different media or platforms, with one copy off-site and outside the blast radius of the primary environment. Many teams now add a fourth condition — one copy immutable or offline — because ransomware that reaches your production credentials usually reaches your backup console too.
The part that gets skipped is restoration. A backup is a receipt; a restore is the product. If you have never timed a full restore of your largest database, you do not know your RTO — you know your aspiration. Time it, write down the number, and compare it to the target you promised operations. The gap is your project list.
For companies whose storefront or customer portal is the revenue channel, this is where an ongoing website maintenance and uptime service earns its keep: monitoring, patching, verified backups, and a documented failover path, handled as routine work instead of a fire drill.
Test the Plan, Staff It, and Keep It Current
An untested plan is a hypothesis. The purpose of testing is not to prove the plan works; it is to find the three things that have quietly broken since the last test.
Run a tabletop first, then a real failover
Start with a tabletop exercise: 90 minutes, the relevant leads in one room, a scenario read aloud, and everyone walks through what they would do. Cheap, low risk, and it reliably surfaces the dependency nobody documented — the vendor portal that only one person can log into, the DNS registrar whose MFA is tied to an ex-employee's phone.
Then schedule a real one. Restore a production database to a staging environment and have a user confirm the data is usable. Once a year, fail over the primary system in a controlled window and keep it running on the secondary for a few hours.
Use a short post-test checklist every time: Did we meet the RTO? Did we meet the RPO? What was missing from the runbook? Who was unreachable? What is the owner and date for each fix?
Who answers the phone at 2 a.m.
Most plans die on staffing. A six-person IT team cannot credibly promise a one-hour RTO around the clock without burning out, and a single on-call engineer with no backup is a single point of failure wearing a pager.
Options worth comparing on cost: a managed service for Tier 1 systems only, a documented degraded-mode procedure that buys hours until business-day staff arrive, or an extended coverage arrangement with a partner. This is one practical argument for a nearshore setup — a Mexico-based engineering team overlaps US business hours in Central and Eastern time zones, so escalation during a West Coast evening or an early East Coast morning reaches someone awake and already familiar with the system, without an overnight rotation in a twelve-hour-offset time zone.
Finally, set a review cadence. Twice a year, plus after any material change: a new system, an office move, a key departure, a shift in regulatory scope. If your recovery plan involves regulated data — healthcare, financial, or cross-border records — confirm the specific retention and notification requirements with legal counsel or a compliance specialist rather than relying on a generic template.
Start With One Page
You do not need a hundred-page document. You need one page per critical system: the RTO, the RPO, where the backups live, who restores them, and how the business operates in the meantime. Three pages that are true beat a hundred that are aspirational.
Write it this quarter, test it next quarter, and revise it after the first test tells you what you got wrong. If you would rather have an engineering partner map your systems, set realistic recovery targets, and build the failover path with you, tell us about your setup and get a quote.
Related articles
Strategy and costs
Disaster Recovery Plan Template for Small Business
Most companies discover their disaster recovery plan is incomplete at 2 a.m., mid-outage, with customers calling. This guide walks through a disaster recovery plan template you can fill in section by section, from RTO and RPO targets to call trees and restore drills. It includes the formulas
· Equipo Strappberry
Strategy and costs
Disaster Recovery Plan Cloud Computing: Practical Guide
Moving to AWS, Azure or Google Cloud does not give you a disaster recovery plan — it gives you better building blocks. This guide walks through setting RTO and RPO targets, choosing between backup-and-restore and warm standby, writing a runbook your team can actually follow at 3 a.m., and testing it
· Equipo Strappberry
Strategy and costs
Testing a Disaster Recovery Plan: A Practical Guide
Most disaster recovery plans are documents nobody has ever executed, which means the first real test happens during the outage. This guide shows how to define recovery targets, pick the right level of test, run your first restore drill in a week, and turn the results into fixes instead of paperwork.
· Equipo Strappberry