How to review offshore work without re-doing it

The most common way an offshore hire fails is not bad work. It is a reviewer who re-performs. They open the file, re-tie the balances themselves, re-check every coded transaction, and end up spending four hours on a job that took the preparer six. The work got done twice and the firm's capacity did not move.

Reviewing and re-performing are different activities. Re-performing means you do the work again and compare. Reviewing means you check whether the work was done (evidence, tie-outs, exceptions handled, conclusions supported) and go deep only where risk is.

Re-performance is a diagnostic tool for a new relationship. It is not a process.

1. Make the work reviewable before you review it

Most review pain is a submission problem. A file that arrives as "the close is done" forces re-performance because there is nothing else to inspect.

Set a submission standard. Nothing enters review without:

• A tie-out. Every balance-sheet account agrees to something outside the ledger (statement, schedule, sub-ledger, confirmation) and the supporting item is attached or referenced where the reviewer will look for it.
• A completed self-review checklist, signed by the preparer. Same checklist every period.
• An open-questions list. Everything the preparer was unsure about, with what they did and why. This list is the most valuable page in the package.
• A variance note. Anything materially different from prior period, explained. Unexplained movement is the cheapest error detector you have.

The self-review checklist saves you more review time than anything else on this list. A competent preparer running their own checklist catches most of what you would have caught, at a fraction of your rate. Write it once against your actual review notes: if you have written a note twice, it becomes a checklist line.

2. Review by risk, not by volume

Depth should follow consequence, not transaction count.

Full review: anything involving judgment or estimate: accruals, cut-off, revenue recognition, capitalization calls, intercompany, new account treatments, related-party items, anything materially unusual. This is where real errors live and where you have no substitute for your own attention.

Targeted testing: accounts that tie but where the tie could be wrong: reconciliations with large unexplained items, aged AR and AP that has stopped moving, clearing and suspense accounts, anything with a manual journal behind it.

Spot-check: high-volume routine coding. Pull a sample, weighted toward large amounts, new vendors, and accounts prone to misclassification. If the sample is clean, stop. If it is not, expand that category, not the whole file.

Skip: routine, tied, low-value, unchanged from a period you already reviewed. Naming what you skip is part of the process, not an admission.

The three-question filter for any account: is it material, does it involve judgment, has it changed. Two yeses means read it properly. Zero means tie-out and move on.

3. Write notes that fix the pattern

A review note that says "reclass this to prepaid" fixes one line. A note that says "prepaid expenses: anything paid this period covering a future period goes to prepaid and amortizes, see the checklist line" fixes the category.

Every note should say what was wrong, what the rule is, and where the rule lives. Notes take slightly longer to write and stop recurring, which is the whole point.

4. Track error patterns by category

Keep a running log. Date, client, account area, error type, whether it was caught by the preparer's own checklist. Ten minutes a month to maintain.

After two or three periods the log tells you things a general sense of "the work is fine" never will:

• Which categories recur: those are training or checklist gaps.
• Which are one-offs: those are noise; do not respond to them with process.
• Whether the self-review checklist is catching things, which tells you if your standard is working.
• Which client files generate the most notes, which is often the file's fault rather than the preparer's.

Recurring notes in one area mean one of three things: the checklist has no line for it, the preparer does not understand the treatment, or the treatment was never written down anywhere. All three are yours to fix, and all three are cheap.

5. Taper depth on evidence, out loud

Trust should be granted on a schedule and stated explicitly, so the preparer knows where they stand and you know what you have committed to.

A pattern that works:

1. Periods one and two: full review, including re-performing selected sections to calibrate what "done" means here. Accept that you gain little capacity in month one. You are buying a baseline.
2. Periods three to five: risk-based review as above. Full depth on judgment, spot-checks on routine. Note volume should fall.
3. After a clean stretch: narrow to judgment areas, variances, and exception reports. Routine accounts move to a quarterly sample.
4. Steady state: you review conclusions and open questions, not the ledger, with a periodic deeper pass to make sure the standard has not drifted.

Two rules make the taper safe. Never taper a category the log says is still producing notes. And re-tighten immediately on a change in scope, a new client type, a new entity, or a new system, not because the preparer got worse, but because the pattern is unfamiliar again.

What this looks like in hours

Say a monthly close takes a preparer 20 hours. Re-performing costs you 12 hours and the firm nets 8. Risk-based review on a properly submitted file costs 3 to 4 hours in the early periods and often under 2 at steady state. Same work, same quality bar, and the difference is capacity you can actually sell.

That gap is the whole reason to hire. It only appears if the submission standard exists and the review discipline holds.

Where the operator model helps

Review tapers faster with a dedicated person than with a shared pool or a ticket queue, for a simple reason: your error log is about one person, so training compounds and your checklist gets shorter over time. Rotate the preparer and you restart at period one. If you are still deciding which parts of the close to hand over, start with how to split close work with offshore staff.

The operators we place work full-time US hours, embedded in your process. Every one has a four-year finance or accounting degree, QuickBooks certification, and role-specific skills testing; most also work in Xero, and they work in Karbon, TaxDome, Sage, Gusto and ADP. Every operator signs a comprehensive NDA before access is granted. The client portal gives you daily reports, live online and offline status, and task visibility, so review is not the only place you find out how the work is going.

If the fit is wrong, the 30-Day Right-Fit Guarantee applies: we re-vet and place someone else at no additional cost, or you exit and we refund the wages you have paid plus a portion of the setup fee. We also run two-week and 30-day check-ins with both you and the operator: the two-week check-in is the right moment to raise a submission standard that is not being met. If the problem is a backlog rather than ongoing capacity, The Books Cleanup Crew is fixed scope and fixed price for up to 12 months of backlog.