Small Business Operations Audit: What to Check and How

Pick a boundary first, or the audit becomes a list of everything that annoys you. Choose one of two shapes. A full-loop audit follows one customer from first inquiry through cash collected. A function audit covers one area, such as hiring or purchasing, in more depth.

For a first audit, follow the full loop. Assign one owner, block a week, and agree on three inputs before starting: the last five completed customer jobs, the org chart with who does what, and the list of software the team uses. Interview the people who do the work, not just their managers. The gap between the process as described and the process as performed is where most findings live.

Work through these areas in order, since each one feeds the next:

• Order to cash: how a request becomes a job, an invoice and a payment, and where it waits between steps.
• Delivery: who does the work, what defines done, and how you catch defects before the customer does.
• People: whether each role has an owner, a backup and written expectations, and how a new hire learns the job.
• Vendors and purchasing: who can commit spend, how approvals work, and whether you know your renewal dates.
• Systems and data: which tool is the source of truth for customers, jobs and money, and where people re-key data.
• Continuity: what happens if the owner, the lead technician or the main system is unavailable for a week.

For each area, write down what you saw, not what you assume. "Invoices go out when the owner remembers" is a finding. "Invoicing needs improvement" isn't.

The walk-through is the most useful hour of the audit. Take a real, recent job and follow it with the people who touched it.

1. Start at the first message from the customer and record who received it and when.
2. Note every handoff, and write the time between the handoff and the next person acting on it.
3. Mark each place where someone re-typed information, searched for it or asked someone else.
4. Record where the work waited, such as for approval, a missing file or a callback.
5. End at the payment and note how many days passed from finished work to cash.

Repeat with a job that went badly. The comparison shows which steps hold up under pressure and which depend on one person's memory.

Score each finding on two questions: how often does it happen, and what does it cost when it does? Cost can be rework hours, delayed payment, customer complaints or exposure to a legal or safety problem. Use high, medium and low for each, and multiply mentally: a frequent, costly problem goes first.

Then separate findings into three buckets: fix this month, schedule this quarter, and accept for now. Accepting is a legitimate outcome. Say a three-person business spends $50 a month on an app: it does not need a purchase-order system for that. Write down why you accepted it, so the next audit doesn't reopen the question.

If your findings lean toward missing or stale documentation, the SOP audit checklist goes deeper on that one area, and the business continuity plan template covers the continuity gaps.

Each fix needs an owner, a date and a definition of done. "Improve onboarding" fails all three. "Owner writes a one-page kickoff checklist for new jobs and uses it on the next three jobs, by the 15th" passes.

Limit yourself to three to five active fixes at once. A small team can't absorb twelve changes, and half will quietly die. Put them in a tracker the team already opens daily. A workflow tool such as Process Street can turn a recurring fix, like a job kickoff checklist, into a repeatable checklist that runs each time.

Common mistakes to avoid:

• Auditing only the areas the owner already suspects.
• Interviewing managers and skipping the people who do the work.
• Producing a long list with no ranking.
• Skipping the follow-up: put a re-check of each fix on the calendar 30 days out.