Most offshore accounting relationships are decided on a sales call, a rate card and a reference the provider chose. Then a twelve-month agreement gets signed, and the firm discovers in month three whether the decision was right.

There is a better way, and it costs roughly the price of one week of a US staff accountant's time.

A structured pilot - forty hours of real work, measured against pre-agreed criteria - tells you more than any amount of due diligence. It reveals what the sales call cannot: whether the person assigned to you can actually do the work, how many of your hours their output consumes, and what happens when something goes wrong.

This is how to design one that produces a decision rather than a vague impression.

Quick answer

A well-designed offshore accounting pilot runs three to four weeks, uses 40 billable hours, covers three to five real client files, and is measured against four criteria: output quality (measured as review points raised per file), turnaround time against commitment, communication responsiveness, and total hours the engagement consumed on your side.

Expect to pay for the pilot. Expect to invest eight to twelve of your own hours in it. And define the pass threshold before the work starts - typically fewer than five substantive review points per file by the final week, with turnaround met on at least 90% of items.

Providers who refuse a paid pilot, or who want to staff it with a senior who will not be your ongoing resource, are telling you something useful at no cost.

Why most pilots fail to produce a decision

Three failure modes account for nearly all inconclusive pilots.

The showcase problem. The provider assigns their best senior accountant to the pilot, then staffs the real engagement with a junior. You tested a person you will never work with again. The fix is contractual: name the pilot resource and require that the same individual continues into the engagement.

The unrepresentative sample. Firms send their cleanest, simplest files to avoid wasting money on a test. The pilot passes, the real work begins, and the first genuinely messy file breaks the process. Send something hard.

No defined criteria. Without a pass threshold agreed in advance, the outcome becomes a feeling. Someone says "it seemed fine, mostly" and the decision defaults to whoever is most persuasive in the room. Write the criteria down before day one.

What to send - and what not to

Send

• One clean, representative file. Establishes the baseline for normal work.
• One genuinely messy file. Unreconciled accounts, missing documentation, inconsistent coding. This is where you learn whether the team applies judgement or just processes.
• One file requiring judgement. Accruals, a lease classification, a multi-entity allocation, revenue cut-off. You are testing whether they know what they do not know.
• One file with ambiguity you have deliberately left unresolved. Do they flag it, guess, or ignore it? This single test predicts more future pain than any other.
• Your actual documentation. Send the process notes as they exist, not a polished version written for the pilot. You need to know whether your real documentation is sufficient.

Do not send

• Your largest or most politically sensitive client.
• Anything under a contractual restriction on offshore processing.
• Tax return information, unless your IRC Section 7216 consents are already in place - consent must precede disclosure, including during a pilot.
• Work with a hard external deadline inside the pilot window.

How to spend the 40 hours

Allocation	Hours	Purpose
Onboarding and knowledge transfer	6–8	System access, walkthrough of your processes, chart of accounts, client context
Clean representative file	8–10	Baseline quality and turnaround
Messy file	10–12	Judgement, exception handling, clean-up approach
Judgement-heavy file	6–8	Technical competence, escalation behaviour
Documentation and handover	4–6	Process notes, open items list, questions log

The onboarding allocation is not overhead - it is part of the test. How efficiently a provider absorbs your process tells you what full ramp will cost.

A four-week pilot plan

Week 1 - Setup and first file

Grant access on day one, not day four. Hold a 60-minute kickoff with the actual assigned accountant present, not only the account manager. Walk through one file live rather than sending written instructions alone.

Issue the first clean file with a stated turnaround commitment. Do not extend it when they ask. How they handle time pressure is data.

Watch for: whether questions arrive early and batched, or late and scattered. Early batched questions indicate someone who read the whole file before starting.

Week 2 - The messy file

Issue the difficult file. Provide no additional support beyond what your normal process would give.

Watch for: Do they attempt to reconcile and flag what they cannot, or return the file with a list of everything they could not do? The first is an accountant. The second is a processor.

Week 3 - Judgement and ambiguity

Issue the judgement-heavy file containing your deliberate ambiguity.

Watch for: the escalation. A good result is a specific question with a proposed answer attached - "the March insurance payment looks like a 12-month prepaid; I have treated it as prepaid and amortised from April, please confirm." A poor result is either silence or an open-ended "what should I do about the insurance?"

Week 4 - Documentation and review

Ask them to produce process documentation for what they learned, plus an open items list. Then run your final review and score the pilot.

Watch for: whether the documentation is usable by a different person. This is your continuity insurance.

The scorecard

Score each criterion 1 to 5 and apply the weights. Agree the weights with the provider before starting.

Criterion	Weight	What you are measuring	Pass indicator
Output quality	35%	Substantive review points raised per file - errors, omissions, misclassifications	Fewer than 5 per file by week 4, trending down week over week
Turnaround	20%	Items delivered by committed date	90% or better, with proactive notice when at risk
Communication	20%	Response time, question quality, proactive flagging	Replies within one business day; questions carry proposed answers
Your time consumed	25%	Total hours you spent reviewing, correcting and managing	Declining week over week; under 20% of delivered hours by week 4

The fourth criterion is the one firms forget and the one that determines economics. A provider billing $15 an hour whose work consumes an hour of your $150 review time for every three hours delivered is not saving you money. Track your own hours from day one.

Reading the trend, not the average

A pilot that starts poorly and improves sharply is a better signal than one that is uniformly adequate. You are buying a relationship that will run for years; the rate of learning matters more than the week-one baseline.

Plot review points per file by week. If the line falls steeply, the team is absorbing your standards. If it is flat, week twelve will look like week one.

What you have to supply

A pilot is a test of the relationship, not just the provider. Your side of the bargain:

• A single named point of contact on your team who will actually respond.
• System access on day one, with the right permission levels.
• Whatever documentation exists, promptly.
• Feedback within two business days of each delivery. Delayed feedback makes improvement impossible and makes the pilot a test of nothing.
• Eight to twelve of your own hours across the four weeks.

Firms that skip the feedback loop consistently report inconclusive pilots. The provider cannot correct what nobody told them was wrong.

Commercial terms worth insisting on

• Pay for the pilot. Free pilots are staffed as loss-leaders and create obligation. Paying keeps both sides honest and gives you standing to demand standards.
• Fixed fee, not hourly. Removes the incentive to spend the full forty hours regardless of need.
• Named resource continuity. Written commitment that the pilot accountant continues into the engagement, with defined notice if they change.
• No auto-conversion. The pilot must not roll into a term agreement by default.
• Data deletion on exit. If you do not proceed, your files must be deleted and the deletion confirmed in writing.
• No exclusivity. Run two pilots in parallel with different providers if the decision matters. The comparison is far more informative than either pilot alone.

Red flags during the pilot

1. The account manager answers every question instead of the accountant. You will never speak to the person doing your work.
2. Review points repeat across weeks. Feedback is not reaching the individual. This is a management problem and it will not improve at scale.
3. Delivery is consistently late but never flagged in advance. Missed deadlines are survivable; unannounced missed deadlines are not.
4. Pushback on the messy file. "This file isn't suitable for outsourcing" usually means "we only do clean work," which describes almost none of your actual portfolio.
5. Pressure to sign before week 4. A provider confident in their output waits for the score.

After the pilot: three possible decisions

Pass - scale deliberately. Do not jump from one resource to five. Add capacity in increments that your review capacity can absorb. The constraint on scaling offshore is almost never the provider's headcount; it is your firm's ability to review.

Marginal - extend rather than abandon. If the trend is improving but the level is not yet there, a second 40-hour block is cheaper than restarting the search. Set a higher threshold for the extension.

Fail - say so specifically. Tell the provider exactly which criterion failed and by how much. You may need them in two years, and the market is small. A precise decline is more useful to both sides than silence.

 

Frequently asked questions

How long should an offshore accounting pilot be? Three to four weeks with roughly 40 billable hours. Shorter than three weeks does not allow a learning curve to appear, which is the single most informative signal. Longer than six weeks delays a decision you already have enough data to make.

Should I pay for a trial with an outsourcing provider? Yes. Paid pilots are staffed properly and give you standing to hold the provider to defined standards. Free trials are frequently staffed by a senior who will not be your ongoing resource, which makes the result unrepresentative.

What should I measure during an offshore accounting pilot? Four things: substantive review points raised per file, percentage of items delivered by the committed date, communication responsiveness and question quality, and the number of your own hours the engagement consumed. The last is the one most firms omit and the one that determines whether the arrangement is economic.

How much does a 40-hour accounting pilot cost? Typically $600 to $1,000 for bookkeeping-level work at standard India rates, and $1,200 to $2,400 where the pilot requires senior or controller-level input. Some providers absorb the onboarding hours.

Can I run pilots with more than one provider at the same time? Yes, and it is advisable when the decision matters. Send comparable work, apply the same scorecard, and tell each provider that a parallel pilot is running. Providers who object to a competitive process are optimising for something other than your outcome.

What if the pilot fails? Determine whether the failure was the provider or the input. Insufficient documentation, delayed feedback and unclear expectations cause more pilot failures than provider capability does. If your side was sound and the output was not, decline specifically and move to the next shortlisted provider.

Do I need Section 7216 consent before a pilot involving tax returns? Yes. The consent requirement attaches to the disclosure, not to the size or purpose of the engagement. A pilot is a disclosure. Obtain consent first, or restrict the pilot to bookkeeping and close work where the requirement does not apply.

 

We would rather be tested than described. Send us three to five real files, your current process documentation and your turnaround expectations, and we will quote a fixed-fee 40-hour pilot with a named accountant who stays with you if you proceed. If the scorecard says no, we will tell you where we fell short.