How to Write a Business Continuity Plan for a Small Business

The disruption that takes a small business down is almost never weather. It is a person: the bookkeeper who gives three weeks of notice, the manager who leaves owning a process nobody else has run, the owner whose doctor tells him to stop. A business continuity plan for a small business is a short written list of the work that cannot stop, the name of whoever picks up each piece when the usual person is gone, and where those instructions live. The first version takes an afternoon.

What is a business continuity plan?

A business continuity plan is a written record of which work must keep happening during a disruption, who does it when the usual person cannot, and where the instructions live. Large companies fill a binder because they have sites, systems, and standby staff. A 30-person company has none of that, so the plan shrinks to a page per process.

The enterprise version documents how to switch to the redundancy. The small business version usually discovers there is none.

That discovery is the real deliverable. Most of what you write down is work somebody performs correctly every week without ever being taught it in writing: the processes you know by heart and have never documented.

The four disruptions a small business actually faces

Continuity planning goes wrong when it opens with catastrophes. In interviews with operators, the interruptions that actually happened fall into four shapes.

Three of the four are people. The fourth is access, and it only hurts where the answer lives in one place.

One staffing operation had its customer service team cut to two people overnight. The urgent problem was not hiring; it was that no training existed for replacements nobody had posted yet.

The ordinary version does more damage because it recurs: what breaks when a key employee calls in sick is the same failure at smaller scale.

For most small businesses the disruption arrives as a resignation letter, not a weather alert.

Start with people, not weather

A founder we interviewed had built three companies and could not take a week away. His team kept telling him that if something happens to you, this goes right down the tube. He had already had health issues. His own account: there are a lot of things that I know that nobody knows but me, and I do not even know what I have told people and what I have not.

Ask an owner in that position what fails first. Customer acquisition, in his case, because he closed every large contract himself, then culture, then the product only he understood.

That order is the skeleton of the plan. The bill, what key person risk costs a business, is paid in the deals that leave with one departure.

Write the plan for the people you cannot replace this month before you write it for the events you have never had.

List the processes that cannot stop for 48 hours

The list stays short when you build it honestly. Forty-eight hours is the useful cut: longer than a bad day, shorter than a vacation.

1. Work from the calendar. Payroll runs Friday, invoices go Monday, the schedule gets built Thursday. Recurring deadlines do not move for your staffing.
2. Ask what a customer feels by Wednesday. Unanswered calls, unscheduled jobs, a quote nobody sent.
3. Add the quiet obligations. Payroll taxes, insurance certificates, license renewals, the vendor who suspends you at 30 days.
4. Stop at ten rows. Ten rows you can staff beats forty you cannot.

One field services owner covered payables himself after a key person left. He found 500 to 800 dollars vanishing every week from skipped audits, missed deductions, and overpayments: up to 10,000 dollars a month, not from theft but from a process nobody followed.

That is the cost of a gap nobody called a disruption. Score each row the way you would find a bus factor of one: how many people could run it tomorrow without a phone call.

A knowledge risk assessment ranks on the same two questions: where knowledge sits in the fewest hands, and where the most damage lands when it breaks.

Write the fallback for each one, with a name on it

A fallback is not a note saying someone in accounting covers this. It is three things: why the process exists, where the judgment calls sit, and how the work gets done click by click. Steps alone survive nothing.

An exiting bookkeeper with three weeks of notice trained her replacement over screen shares, because none of the payroll and commission process was written down. Describing one weekly run took 85 minutes: reports merged by hand because the export strips the job IDs, screenshots of group chats to identify the helper, receipts hunted through a login she did not have. Every workaround was a decision, not a step.

Four rules turn a name into a fallback that holds.

1. Put a person on it, not a role. Roles do not answer the phone at 7am, and the name has to know it is theirs.
2. Record the real run. Capture the practitioner working on the day it happens, screen shared and narrated. Budget 2 to 4 hours per process.
3. Write down the forks. Every process has moments where the right answer depends on context. Handovers break there.
4. Date the page. A date tells the next reader how much to trust it.

At a staffing firm, the taxonomy for escalated cases (injury, altercation, theft, damage) left with the manager who owned it, and afterward everyone logged the same cases differently. The steps had not changed. The judgment had.

A fallback with no name is a hope, and a name with no recording is a favor asked at the worst possible moment.

Done with runway instead of a resignation letter on your desk, this is the same work spread over months: a knowledge transfer plan built before you need one.

---

Where does the plan live when the office is unreachable?

Somewhere every fallback person can open from a phone: without the office, without your network, without asking you for a password. That rules out a binder on a shelf, a laptop folder, and the shared drive nobody opens.

One of our own client demos ran three days into a home power outage, from someone else's house, on schedule, because nothing it needed sat inside the dark building. That is the whole test for storage.

If the covering person cannot find the instructions alone in five minutes, the storage has failed.

Test it with a planned absence

You test a continuity plan by removing a person on purpose and watching what happens. Not a tabletop exercise. An absence on the calendar.

An electrical contractor we worked with planned two weeks away and called it a measurement tool, not a vacation: you may get some honest feedback when I am not expected to pop into your office. No meeting produces that data.

The test has four parts, and the last one matters most.

1. Announce the dates, not the test. People perform for a drill. They behave normally for a vacation.
2. Close the escape hatch. Every question routes to the fallback name first, even the nine-second ones.
3. Log what got asked anyway. Whoever covers notes what they had to ask, look up, or guess.
4. Edit the plan from the log. Each line is a gap with a location attached.

The measure is not whether the two weeks felt smooth. Smooth usually means everyone waited for you to come back.

An absence you scheduled is the only continuity test that costs nothing and tells the truth.

The log is the fastest route to a business that runs without you: every line names a decision with exactly one owner.

Review the plan whenever the org chart changes

Annual review dates are how continuity plans go stale. The triggers are events: a hire, a departure, a promotion, a new service line, a software change.

That last one catches people. One ops leader rebuilds his reporting pipeline every six weeks or so, because a renamed campaign snaps the key tying records together. Any fallback that says run the weekly report inherits that fragility.

Tie the review to the event: when a name on the plan changes jobs, the plan changes that week.

The Systems Effect does this work for a living, interviewing the people who hold the undocumented parts of a business and turning what they say into SOPs, training and process maps.

You do not need the whole plan this month. Pick the person whose two-day absence you would feel by Wednesday, write the three things their fallback needs, and put a second name on the page.

Frequently Asked Questions

What is a business continuity plan for a small business?

A business continuity plan is a written list of the work that cannot stop, the person who takes over when the usual one is unavailable, and where the instructions live. In a small company it runs to a page per process, not a binder, and answers one question fast: who does this now.

What should a continuity plan include?

Include five things for every process that cannot pause for 48 hours: the process, who normally runs it, the named fallback, where the instructions live, and the date somebody last proved it works. Attach the purpose and the judgment calls, not just the steps. A row with no name and no date is a wish.

How do you test a business continuity plan?

Remove a person on purpose for a scheduled stretch, ideally two weeks, and let the fallback run the work. Log every question the covering person had to ask anyway, because that log is your edit list. A tabletop meeting tests the document; a planned absence tests the business.

How is continuity planning different from disaster recovery?

Disaster recovery restores systems and data: backups, hardware, accounts, and how fast you get them back. Continuity planning keeps the work moving while that happens, which in a small business is mostly a people question. Under 50 people, the people half fails more often.

How often should a continuity plan be updated?

Update a continuity plan on events rather than a calendar: a hire, a departure, a promotion, or a tool change that alters how a process runs. A yearly review is a backstop, but the plan goes stale the day a named fallback changes jobs.