What a CPA costs a small business depends on what you're buying: tax preparation, bookkeeping, full-service accounting, or hourly advisory work. Fees vary with entity type, transaction volume, and the condition of your books, and more firms now charge flat fees instead of billing by the hour. DIY tax software costs less but won't catch strategic mistakes. Most CPA fees pay for compliance, not optimization. A once-a-year preparer can document your decisions but can't change them, so the real cost of a CPA is often the tax savings a compliance-only relationship never pursues.
Ask five small business owners what their CPA costs and you will get five wildly different answers, because they are buying five different things. One is paying for a tax return. One is paying for a tax return plus twelve months of bookkeeping. One is paying for someone who calls in October to tell them to do something before December 31.
Those are not the same product, and comparing their prices tells you nothing. Here is an estimate of what each piece actually costs based on published data, along with a note on how stale some of the numbers you will find elsewhere really are.
For a small business with reasonably clean books:
Now, the details and the caveats that matter.
A business tax return averages $654 to $775 per form, but ongoing monthly service is where most of the cost sits.
The per-form fee is the smallest number here. What drives your total cost is whether you pay for a return once a year or for bookkeeping and planning every month.
Most articles on this topic cite the National Society of Accountants' Income and Fees Survey. Worth knowing before you trust those numbers: the most recent completed NSA survey covers data from 2020 and 2021. It is five years old. Fees have increased considerably since then, and most preparers raise their fees every year or two.
For reference, the NSA 2020-2021 survey reported average fees of $323 for a Form 1040 with Schedule A plus a state return, $220 for a non-itemized 1040 plus state, $192 to add a Schedule C, $903 for a Form 1120S, $733 for a Form 1065, and $913 for a Form 1120.
The more current benchmark is the Cornerstone Report, reporting 2025 fee data published in early 2026:
Note the trajectory on that base 1040 figure. A 46% increase across two years is not a rounding error, and it is the clearest signal that any pre-2024 fee data you find should be treated as historical rather than current.
Separately, Ignition's 2025 pricing benchmark found that the most common price bracket for a business tax return was $1,000 to $1,499 (29% of firms), followed by $1,500 to $1,999 (26% of firms). That is higher than the per-form averages above, which makes sense: real engagement is rarely a single form in isolation.
There is no authoritative national survey of CPA billing rates. The figures circulating online, generally $150 to $500 an hour for a CPA and $20 to $80 an hour for a bookkeeper, come from firms' marketing content rather than from controlled research. Treat them as a directional range, not a benchmark.
One useful anchor is a firm that publishes its actual rate card. WCG CPAs & Advisors, updated April 2026, lists rates of $150 per hour for staff accountants, $250 for supervisors, $300 for managers, $400 for partners, and $475 to $525 for senior partners, with tax resolution work at $375 per hour. That is one firm in one market, but it is a real published number rather than an estimate.
Do not confuse billing rates with what accountants earn. The Bureau of Labor Statistics reports that the median pay for accountants and auditors is $81,680 per year, or $39.27 per hour, with 5% projected employment growth from 2024 to 2034. The gap between $39 an hour in wages and $300 an hour in billing covers overhead, software, insurance, unbilled time, and firm profit. It is not markup for its own sake, but it does explain why hourly engagements get expensive quickly.
Current 2026 pricing data from an outsourced accounting cost analysis:
The bundled versus unbundled distinction matters more than the headline number. Paying $500 a month for bookkeeping and then $1,500 in March for a tax return is roughly $7,500 a year, and the person filing the return is often seeing your books for the first time in nine months. That gap is where planning opportunities go to die.
Your last return is the fastest way to find out what a compliance-only relationship cost you. Formations clients save an average of $14,801 a year with a setup built around planning rather than filing.
Six variables explain most of the spread between quotes.
Entity type. A Schedule C sole proprietorship is the cheapest return to prepare. An S-Corp or a partnership costs meaningfully more, and a C-Corp even more. The Cornerstone data above clearly shows the step function.
Transaction volume. More activity means more reconciliation hours, regardless of revenue. A business with $200,000 in revenue across 4,000 small transactions can cost more to service than one with $2 million across 40 invoices.
Book cleanliness. This is the single biggest controllable factor. Firms routinely price disorganized records at two to three times clean-book rates because cleanup is unbilled discovery work until it is not. Consistent monthly bookkeeping is not just hygiene; it is a discount.
Multi-state exposure. Every additional state filing adds cost, and the trigger is often lower than owners expect once remote work and nexus rules are involved.
Timing. Firms are capacity-constrained from February through April. Rush and late work carry surcharges at most firms, and the best ones simply stop accepting new work in March.
Advisory versus compliance. Ignition's benchmark found tax planning and advisory engagements clustering above $2,000 a year, well above compliance-only tax prep. That is a different service, and the pricing reflects it.
One more thing worth planning for: 80% of firms in the Ignition benchmark planned fee increases for 2026, with 30% raising by about 10% and 37% by about 5%. If your fee has not moved in three years, it is probably about to.
The industry has moved away from the billable-hour model, and it is worth knowing where the market stands before you accept an hourly quote.
Per Ignition's 2025 benchmark, the fixed-fee model is now the dominant model for tax preparation, at 37% of firms, up from 35% the prior year. "Minimum fee plus complexity" fell to 23% from 30%. Value pricing is held at 19%. For tax planning and advisory work, fixed-fee and value pricing are tied at 30% each.
What that means for you as a buyer: a flat fee or monthly subscription is not an unusual ask. It is the direction the profession is already moving, and it aligns incentives better. Under hourly billing, every question you ask costs money, so you ask fewer, which is the opposite of what a good advisory relationship should produce.
Worth having the alternative priced out, because it is the real decision most solopreneurs are making.
Current published prices as of August 2026:
So a self-employed person can technically file for under $200 and keep books for $240 a year. Against a $723 S-Corp return plus $6,000 in annual bookkeeping, DIY looks like it wins by thousands.
It wins on price. Whether it wins overall depends entirely on whether the software would have caught what a person would have caught, and software does not tell you that your entity is wrong, your salary is set badly, or that you should have opened a solo 401(k) in October.
Every price above is a compliance cost. None of them measures what actually decides whether the relationship was worth it: the savings you didn't get. That's opportunity cost, and it never shows up on an invoice.
There is no credible study measuring how much small businesses lose through compliance-only accounting relationships. Plenty of firms assert a number, but none cite research, so be skeptical of any figure presented as if it came from a study.
The closest federal evidence is old but telling. The GAO estimated that on as many as 2.2 million tax year 1998 returns, taxpayers took the standard deduction when itemizing would have saved them money. They overpaid by as much as $945 million, an average of $438 per return. About half of those returns were prepared by a third party, mostly paid preparers. That study measured one missed decision on individual returns a quarter century ago. The dollar figure matters less than the pattern: having a preparer did not prevent the overpayment.
The GAO also reported in February 2026 that more than half of individual taxpayers used a paid preparer during fiscal year 2024. Using a preparer is normal. Using one who only shows up at filing time is also normal, and that's the part worth examining. Formations publishes the patterns it sees most often when reviewing prior returns in Your Tax Return Is Likely Leaving Money on the Table.
Take a single example. An S-Corp owner in the 24% federal bracket who makes the maximum 2026 employee 401(k) deferral of $24,500 defers $5,880 in federal income tax that year, before state tax. For an S-Corp owner, that deferral has to come out of W-2 wages through payroll during the year. A preparer who first sees your file in March can report that it didn't happen; they can't make it happen.
The same is true of entity election, reasonable salary, retirement contributions, equipment timing and estimated payments. If nobody was involved during the year, the return is an accurate report of what happened, and that's all it can be. The cost also compounds: a missed decision is usually missed again every year the arrangement stays the same.
Compliance means the return is accurate and defensible. Optimization means the return reflects the lowest tax the law and your facts support. A return can be flawless on the first and weak on the second.
Most CPA relationships are built for compliance, and the incentives explain why. Under IRC Section 6694, a preparer who signs a return with a position that lacks substantial authority can be penalized the greater of $1,000 or 50% of their fee for that return. For willful or reckless conduct, it rises to the greater of $5,000 or 75%. There is no matching penalty for a return that overpays. When the downside of an aggressive call lands on the preparer and the downside of a conservative one lands on you, the rational choice for the preparer is conservative.
In practice, that often shows up as a reflexive no. The Augusta Rule, accountable plans, retirement stacking and cash balance plans are all written into the tax code. Each one also takes documentation, deadlines and follow-through during the year. A compliance-oriented CPA may turn them down as "not worth the risk" when the real issue is that they don't fit a once-a-year engagement. Optimization takes more effort, and effort is usually what a compliance engagement leaves out. The most common version of this is covered in Why Your CPA Says "No" to Your S-Corp.
Caution isn't a flaw. A CPA who refuses aggressive positions is protecting you. The useful question is whether a "no" reflects the law or the scope of the engagement. A legal strategy with clear authority and good documentation isn't a risk; it's work.
Four questions that surface more than a price comparison will.
"How many times will we talk between April and December?" If the answer is zero, you are buying compliance, and you should price it like compliance.
"Is this a fixed fee, and what is explicitly out of scope?" Scope creep is when flat-fee engagements turn into hourly ones.
"Who does the actual work?" A partner rate on the engagement letter and a first-year staff accountant doing the return are common and not inherently wrong, but you should know.
"What did you find last year that saved me money?" This is the only question that separates a filer from an advisor, and the answer is usually immediate and specific or vague and uncomfortable.
"What authority does this strategy lack?" When a CPA declines a strategy, this separates "it isn't allowed" from "it isn't in scope." A specific citation is a real answer. "We just don't do that" is a scope answer.
*Cornerstone figures are average fees for 2025 (published early 2026). Ignition figures are from its 2025 benchmark. Bookkeeping ranges are from inDinero's 2026 analysis. The WCG rate card is dated April 2026, and software prices are as of August 2026. These are market averages, not quotes; entity type, book condition, and state filings move actual prices. QuickBooks prices as of September 2026.