How do I make a small-business continuity plan?

Make a small-business continuity plan by identifying the work that cannot wait, the dependencies that could interrupt it and a funded alternative for a defined disruption. Assign activation authority, customer communications and a return-to-normal check. Then rehearse the plan against the resources people can actually reach.

Begin with one credible interruption. “Keep the business running” is too large to test. “Handle Wednesday's accepted customer work if the office is unavailable for forty-eight hours” gives the team something concrete to examine.

Establish the situation before starting the workaround

Confirm that people are safe and follow applicable emergency instructions before treating the event as a scheduling problem. The person handling customer work cannot declare a building safe, restore utility service or require colleagues to travel through an unsafe area.

This guide's fictional case assumes an office is temporarily inaccessible from Monday at 9 a.m. until Wednesday at 9 a.m. Staff are safe, transport is available and the interruption does not affect the proposed alternate workspace. There is no suspected compromise of the business's systems or credentials. A different situation needs a different response.

The SBA's overview of its Business Resilience Guide identifies essential operations, dependencies, key partners, vital resources and financial readiness as planning areas. Those categories connect the plan to the business rather than to a folder labeled “emergency.”

North Carolina Emergency Management's business guidance also highlights updated employee contacts, operating-status communications and alternative arrangements for incoming or outgoing shipments. Use the relevant local emergency information for the actual event; a general example cannot determine conditions where your team is.

Choose what must continue and what can pause

List accepted customer obligations, time-sensitive payments and the minimum internal work required to support them. Give each a deadline, the consequence of delay and the person responsible. Do not assume every ordinary activity has equal priority.

Our fictional catalog-formatting business has six accepted updates due Wednesday. Five are due by 5 p.m.; a sixth has a confirmed noon deadline. Each requires two paid production-and-checking hours, using approved customer material, and earns $300 when delivered as agreed. There are no deposits, other jobs or late-payment assumptions in this finite case.

The company can postpone speculative marketing and an internal presentation. It cannot quietly postpone the six deliveries or stop communicating with their customers. A manager might approve spending, but only an actual agreed change can alter the customer's delivery expectation.

Write the minimum service level precisely: deliver the six specified updates through the established authorized route, preserve their current approved versions and keep customers informed if an obligation becomes uncertain. The objective is not to reproduce every feature of the office for two days.

Find the dependency beneath the dependency

The work needs people, suitable devices, approved files, legitimate access, a usable workspace, connectivity, checking time and the delivery route. “Work remotely” does not establish any of those.

Ask whether two apparent alternatives share a point of failure. A second laptop inside the closed office is unavailable. A second workspace relying on the same interrupted utility may be unavailable too. A contact list accessible only through the account you cannot enter is not a usable contact list for that scenario.

In this case, two employees have their already-approved company devices with them and can use the established secure work process. The company's cloud service and access controls remain available. An alternate location has confirmed two suitable private seats, power and connectivity for Monday and Tuesday. These are supplied conditions of the fiction, not resources the author has tested or providers being recommended.

The plan still requires the employees to verify access to the actual approved work before reporting that production has resumed. Having a login screen load is not the same as having the correct files, versions and delivery permissions.

A bounded operating record for this case

Scroll across to follow the dependencies and decision points. Use the arrow keys when the table area is focused.

Pay for the ordinary work and the workaround

The two employees have twenty paid hours scheduled across Monday and Tuesday at $30 an hour, totaling $600. The plan assigns twelve hours to the six jobs, four to communications and confirming requirements, two to secure workspace setup and two to recording work and payments. Those twenty hours include all required work during the two days.

Four additional paid staff hours on Wednesday cover recovery and reconciliation, costing $120. They are available without displacing another accepted job. The business is not asking someone to perform an extra unpaid evening of recovery after declaring the incident over.

The two-day workspace costs $300 in total, including both seats and the stated facilities. Transport costs $80 and incremental remote-access support $40. Continuing overhead for Monday through Wednesday is $180. There are no other payments, equipment purchases, depreciation, taxes or financing charges in this finite illustration.

The complete cost is therefore $1,320: $600 ordinary wages, $120 recovery work, $420 for the alternate arrangement and $180 continuing overhead. Of that, $540 is additional to the $780 of ordinary wages and overhead. Showing only the extra $540 would omit bills the company must still pay.

The business has $2,000 available at Monday's start, separately from other obligations. It pays $420 for workspace, transport and support on Monday, $600 wages on Tuesday and $300 for recovery and overhead on Wednesday. Cash falls to $680 before the customers' $1,800 arrives on Friday. It then reaches $2,480.

The chosen reserve is $500, leaving only $180 above that floor at the low point. No insurance payment, emergency loan or customer acceleration is assumed. A possible reimbursement is not spendable money until its amount and timing are established.

With only $1,000 available instead, the same plan would be $320 short before collections, even before preserving a reserve. The manager must address that gap before authorizing the workaround. An available workspace does not fund payroll.

Say who can activate what, and until when

For this scenario, the operations lead can authorize the stated $540 of extra continuity spending and direct the scheduled work. The permission covers two workspace days and the four recovery hours. It does not cover a third rental day, a replacement system or a changed customer obligation.

By 11 a.m. Monday, the lead must know whether both employees can reach the current briefs and produce through the authorized route. This is a decision checkpoint chosen for the case, not a universal recovery target. It leaves time to contact customers while alternatives remain possible.

If access fails, the plan has reached an unresolved condition. The lead records the failure, preserves legitimate available records and assesses what can still be delivered. It does not ask employees to move customer files into personal accounts or bypass access controls to keep the forecast intact.

A useful update states what is known, what remains uncertain and when the next confirmed update will arrive. Avoid “business as usual” while the team is still checking whether the usual work is possible.

Check whether the backup describes today's promises

A continuity copy can be complete for its date and still omit newer commitments. The fictional company has a Sunday 6 p.m. export of its six-job register. At 8 a.m. Monday, customer J2 approved a new input version and customer J5's noon deadline was confirmed. The export predates both changes.

The current authorized system and its change history remain accessible in this scenario. The employees compare that history with the export, identify the two updates and record the current version and deadline before assigning work. They do not overwrite the live register with the older copy.

That detail changes the rehearsal. If you hand someone only the Sunday export and ask them to prioritize the work, can they tell it is incomplete? If the answer depends on a colleague remembering a call, the continuity record needs a repair.

This is a paper exercise, not a claim that a backup has been restored or tested. A real restoration test needs the appropriate authorized people and a safe procedure that protects production. In a suspected security incident, involve the responsible technical and incident-response specialists rather than assuming this office-access scenario applies.

Treat the return as part of continuity

When the office is available on Wednesday, do not simply switch every routine back on. Check which jobs were completed, which files are current, which messages were sent and which payments were made. A second dispatch or duplicate charge can be created during a very enthusiastic return to normal.

The four funded recovery hours cover that reconciliation and the handback. Each job keeps one owner. The lead confirms that the delivery record matches what actually reached the customer and that temporary work has been incorporated into the ordinary process.

The founder-absence guide focuses on whether information and authority remain usable without the founder. The incident-cost guide follows the whole commitment as a workaround grows. This plan needs both habits, but adds a defined interruption, prioritized service, resource dependencies and a controlled return.

If the closure extends beyond Wednesday morning, the two-day plan has expired. The lead must establish availability, staffing, further payments and affected customer dates before presenting an extension as secured. An emergency label is not an unlimited purchasing authorization.

Rehearse one failure and repair what it reveals

1. Choose a specific interruption. Name what is unavailable, the period being planned and conditions the scenario assumes remain safe and usable.
2. Rank the obligations. List deadlines and consequences, then identify work that can genuinely pause.
3. Confirm dependencies. Check people, permission, current records, workspace, tools and delivery routes.
4. Fund the period. Include ordinary bills, extra costs and the lowest cash point before collections.
5. Set activation and escalation points. Name who can commit resources, the limit and what happens when a prerequisite fails.
6. Reconcile the return. Bring versions, jobs, communications and payments back together before closing the incident.

Change one condition in the rehearsal: one employee unavailable, access failing or the closure lasting longer. Preserve the original plan and record the gap. Then assign a concrete repair rather than congratulating the team for improvising around it.

Infinite Upside keeps the work behind a customer promise in view. It is a useful wider read when an owner is deciding what the operation must genuinely be able to carry. Read the opening for the ownership story, and use your continuity plan to make the next difficult week less dependent on a last-minute rescue.