Short answer

A compilation, a review, an audit, and a quality of earnings review test four different things, but all four draw on the same underlying accounting records. The engagement you choose changes who signs the cover page and what they conclude. It does not change the monthly close, the reconciliations, and the supporting documentation that every one of them pulls from.

Founders usually meet these four reports as a shopping decision. Which one do I need, what does it cost, who do I call. That is a reasonable question and it is not the one that decides how the engagement goes.

The question that decides how it goes is whether your records can answer what the engagement asks. A firm that has to build your general ledger detail before it can test anything is going to be slower and more expensive than the same firm working from a clean close, no matter which of the four you hired them for.

The four engagements, compared by what they demand from your records

Dimension	Compilation	Review	Audit	Quality of earnings review
What it concludes	Nothing. It presents management's numbers in financial statement format	Under AICPA standards, limited assurance that nothing came to the accountant's attention indicating material misstatement	Under AICPA standards, an opinion on whether the statements are fairly stated under the applicable framework	Not a standardized engagement, so the conclusion varies by provider; typically a normalized earnings analysis and an assessment of how repeatable it is
Who performs it	A CPA	A CPA, independent of the company	A CPA, independent of the company	A transaction advisory or accounting firm; independence requirements differ from the assurance engagements
What it asks of your books	Numbers, in whatever condition they exist, plus a trial balance	Explanations. Analytical procedures generate questions about variances, and someone has to answer them from the records	Evidence. Balances get tested against third-party support, so reconciliations and source documents have to exist	Detail. Transaction-level revenue data, monthly rather than annual figures, and documentation behind every adjustment
Relative cost and duration	Lowest of the four, shortest	Higher than a compilation	Highest of the assurance engagements, and longest in a first year	Varies widely by scope; drawn out sharply when records are incomplete
What it does not fix	Anything. It reformats what you give it	Underlying record quality; the procedures are limited by design	Whether your adjusted earnings are supportable, which is outside the scope of an audit opinion	The close itself. It analyzes records, it does not rebuild them
Where founders lose the most time	Assembling a trial balance that ties	Answering variance questions without monthly detail to answer from	Producing third-party support for balances that were never reconciled	Reconstructing adjustments from memory instead of documentation

Read one column at a time and the pattern is easier to see.

A compilation is a presentation service. A CPA takes management's numbers and formats them into financial statements without testing them and without expressing assurance. It asks the least of your records and returns the least, which makes it useful for internal discipline and largely silent in any external conversation.

A review is a limited-assurance engagement under AICPA standards, built on analytical procedures and inquiries rather than testing. What it asks of you is explanations. When a margin moves and the accountant asks why, the answer has to come from your records, and a founder without monthly detail ends up answering from memory.

An audit is the highest level of assurance a CPA provides on historical statements, and under AICPA standards it ends in an opinion on whether those statements are fairly stated under the applicable framework. What it asks of you is evidence: bank reconciliations, confirmations, invoices, contracts, and support for balances. Records that were never reconciled do not become reconciled because an auditor arrived. Whether a founder needs one at all is a separate question, and we took that one on in why clean books beat an audit when you sell.

A quality of earnings review is a diligence exercise that analyzes whether reported earnings are accurate, normalized for one-time and owner-specific items, and likely to repeat. It is not a standardized engagement, so its scope and depth vary meaningfully between providers. What it asks of you is detail, at a monthly and transaction level, plus documentation behind every adjustment claimed.

Why the four sit on top of the same accounting work

All four engagements read the same records, which is why record quality drives the outcome of every one of them. The trial balance, the reconciliations, the general ledger detail, and the supporting documents are not engagement-specific. They are the inputs.

That is also why the sequencing question matters more than the selection question. Hiring an assurance engagement to compensate for an incomplete close converts an accounting problem into a professional-fees problem. The work still has to happen. It just happens on someone else's hourly rate, under a deadline, with a third party watching.

The reporting standard is the same one a buyer's diligence team applies, which is why the nine reports in can your books answer what buyers ask overlap heavily with what these engagements request. Build once, use in both settings.

The shared prep list: the artifacts every one of the four pulls from

Before any of these four engagements starts, eight artifacts do most of the work. This is the distinctive part of the answer, because the list does not change with the engagement you pick.

A closed month, every month, on an accrual basis. Revenue recognized when earned and expenses when incurred, with the close finished on a predictable calendar rather than reconstructed at year end. Whether accrual is the right basis for a specific business depends on its size, industry, and reporting obligations, so confirm the basis with your accountant rather than assuming.
Bank and credit card reconciliations, completed monthly and retained. Unreconciled cash is the single fastest way to turn any engagement into a cleanup project.
A trial balance that ties to the financial statements you hand out. If the statements circulating internally do not agree with the trial balance, every engagement starts with a reconciliation nobody scoped.
A reconciliation from your internal statements to your filed tax returns. Differences are normal and expected. Unexplained differences are the problem, and explaining them is a records exercise, not a tax exercise.
Revenue detail by customer and by month, retained for multiple years. Annual totals cannot answer questions about seasonality, mix, or concentration, and all four engagements eventually ask one of those questions.
Gross margin by month and by line of business. Margin questions are the most common thing an analytical procedure surfaces, and they are unanswerable from a single annual gross profit figure.
A documented adjustments schedule, built as adjustments occur. One line per item, with the invoice, agreement, or payroll record attached at the time. Adjustments assembled from memory later get discounted, because there is nothing behind them.
A fixed asset and inventory listing that agrees to the balance sheet. These are the balances most likely to be tested against something physical, and the ones most likely to have drifted quietly.

Build those eight and every engagement on the list above gets shorter. Skip them and the engagement bills you for building them.

What to do before you pick an engagement at all

The order is close first, documentation second, engagement third. A founder who reverses that order pays for the same work twice, once as an accounting cleanup and once as professional fees on top of it.

Start with the close, because it is the input to everything else and the only item on the list that compounds. A business that closes cleanly for twelve consecutive months arrives at any engagement with a track record rather than a reconstruction. Then build the documentation as it happens, not in a sprint. Then, and only then, have the conversation about which engagement your situation actually calls for, and let the party who will rely on the report drive that answer. When the question is which report a buyer or lender will lean on and what it does for a transaction, that is an M&A question, and Texas Exit Advisors covers what a quality of earnings report does in a deal.

Two things are worth repeating because they are where founders lose the most money. None of the four engagements repairs your accounting; they report on it, in different words and at different depths. And the prep list is shared, so the work you do now is not wasted if the engagement you eventually need turns out to be a different one than you expected.

Last reviewed: August 2026. This is general information, not accounting, tax, or legal advice. Engagement scope, independence requirements, and the applicable reporting framework vary by company, industry, and the party requesting the report. Thryve Accounting & Advisory is not a tax preparation firm and does not provide income tax planning or filing.

Where Thryve fits

Thryve Accounting & Advisory builds the part nobody can outsource to an engagement letter: a monthly close that actually closes, reconciliations that stay current, adjustment documentation captured as it happens, and reporting detailed enough to answer the questions an accountant, a lender, or a buyer's diligence team will ask. That is the work that makes every one of these four reports cheaper, faster, and less stressful.

If you are looking at a compilation, a review, an audit, or a quality of earnings review and you are not confident your records can carry it, start with a conversation about the close. It is the cheapest place to fix this and the only one that pays off no matter which engagement you land on.

What Each Financial Report Asks of Your Books