How do you start a bookkeeping firm in 2027?

The outcome you should expect

A solo virtual bookkeeper who executes this plan competently lands somewhere between 15 and 40 recurring clients in the first twelve to eighteen months. At the market-median monthly fee — roughly $650 for a client under $1M in revenue, closer to $1,800 for a client in the $1M–$5M band — that book produces $6,000 to $36,000 a month in recurring revenue. Gross margin on compliance-only bookkeeping runs 55–70% before owner compensation, and because a home-based virtual firm carries almost no fixed overhead beyond software seats and insurance, owner take-home lands in the $50K–$280K range depending on how far up the client-size ladder you climb and how quickly you learn to say no to bad-fit work.

That is the honest base case, and it is worth stating plainly because the category is full of inflated promises. The firm that adds staff and moves into Client Accounting Services — the bundled model where bookkeeping, bill-pay, payroll oversight, sales tax filing, and a monthly advisory call ship as one subscription — reaches a different altitude. CAS engagements carry a median around $3,200/mo, with top-quartile firms billing $6,000–$15,000 per client per month. A firm with one or two staff serving 25–60 clients at $1,500–$5,000/mo is a $450K–$3.6M revenue business inside two to three years, though gross margin compresses to 45–65% because you are now paying people.

The timing matters more than the tactics. Three things converged that did not exist a decade ago. First, AI-driven transaction categorization inside QuickBooks Online, Xero, and the automation layer around them (Booke AI, Truewind, Vic.ai, Digits) now classifies the large majority of routine transactions accurately, and receipt-capture tools like Hubdoc, Dext, and AutoEntry have removed most manual data entry. Bank reconciliation labor that consumed roughly forty hours a month per $1M client in 2018 now runs eight to fifteen hours on a modern stack. Second, the CPA pipeline shrank: US accounting bachelor's degrees fell about 17% between 2016 and 2022, CPA exam candidates fell roughly a third, and a large majority of practicing CPAs are within retirement range. CPA firms are pushing compliance bookkeeping out the door to protect partner hours for tax and audit. Third, remote-first practice became normal, which erased the $30K–$120K annual office cost that used to gate entry.

Net effect: a solo operator with a good stack can profitably serve 30–60 small clients where the same person in 2014 topped out around 10–18. Per-client revenue requirements drop, per-firm capacity rises several-fold, and the constraint shifts from processing labor to client acquisition and judgment. That is a very different business than the one the "shoebox of receipts at $50/hour" playbook describes, and if you build for the old constraint you will build the wrong firm.

One caution on expectations. The revenue ceiling for a generalist practice — "bookkeeping for small business," no niche — sits stubbornly around $200K–$500K, because acquisition cost per client never falls. You are always explaining what you do from scratch, always competing on price with the person down the road and with Intuit's own direct-to-consumer bookkeeping service at the bottom of the market. Every meaningful lever described below assumes you have chosen a lane.

What drives that outcome

Four inputs determine which end of those ranges you land on, and they compound rather than add.

Niche choice. Vertical specialization is the highest-leverage single decision. Niche-specialized firms consistently report meaningfully higher per-client revenue, materially lower acquisition cost, better gross margin, and faster organic growth than generalists. The mechanism is straightforward: "bookkeeper for real estate investors" or "Shopify seller bookkeeping" converts dramatically better than "bookkeeper in Cleveland," because the searcher is describing a problem, not a location. Verticals that consistently support premium pricing include real estate investors and property managers (Schedule E, 1031 exchanges, cost segregation, depreciation strategy), e-commerce and Amazon FBA sellers (multi-state sales tax, inventory and COGS reconciliation), restaurants (tip reporting, prime cost, food-cost KPIs), dental and healthcare practices (production versus collection, insurance reconciliation), law firms (IOLTA trust accounting, which carries genuine bar-compliance risk and prices accordingly), construction and trades (job costing, WIP schedules, lien waivers), SaaS and venture-backed startups (deferred revenue, ARR reporting, investor-ready financials, R&D credit coordination), nonprofits (fund accounting, Form 990 support, grant compliance), trucking (IFTA fuel tax, driver settlements, factoring reconciliation), and cannabis, where IRS Section 280E expense restrictions create expertise scarcity and pricing power well above generic work. Typical monthly fees in these lanes run $1,200–$5,000, with SaaS and cannabis frequently higher.

Pricing model. Hourly billing caps revenue at owner-hours, invites collection arguments, and punishes you for getting faster — which is perverse in a period when automation is making everyone faster. Fixed monthly subscription pricing with defined tiers is the structural fix. A workable ladder: Bronze at $400–$900/mo for monthly close, reconciliation, and financial statements; Silver at $1,200–$3,000/mo adding bill-pay, AR follow-up, payroll oversight, sales tax filing, and a quarterly call; Gold at $3,500–$9,000/mo adding a monthly advisory call, KPI dashboard, thirteen-week cash flow, and budget-versus-actual; Platinum at $8,000–$20,000/mo for genuine fractional CFO work — board pack, investor reporting, modeling, capital strategy. Scope has to be written down in the engagement letter, with out-of-scope work billed separately at $150–$300/hour or as a fixed-fee project. Reprice the book 3–5% annually on each client's anniversary; firms that skip this quietly lose a fifth of their real pricing over five years.

Delivery cost structure. The gap between a 30% EBITDA firm and a 12% EBITDA firm is usually labor mix, not revenue. Firms that scale use a three-layer model: automation handles routine categorization and reconciliation, junior or offshore staff handle transaction processing and document chasing at roughly $8–$25/hour through providers like QXAS, Entigrity, or Philippines-based teams, and US-based senior staff handle review, client relationship, and advisory at $80–$200/hour. Blended cost of delivery lands around $25–$45/hour versus $75–$120 for an all-domestic team. This is not a moral question but it is a management one — offshore delivery only works with documented workflows, a real reviewer layer, and clear client disclosure, and firms that skip those three things generate rework that erases the savings.

Advisory attachment. Compliance bookkeeping is a 50–60% margin product with real churn. Advisory sits on top at 60–80% margin and cuts churn substantially, because a client who is on your monthly call discussing cash position does not shop your invoice. The sequencing is what most operators get wrong: you cannot lead with advisory. Six to twelve months of flawless, boring, on-time books earns the data and trust that make an advisory conversation credible.

Benchmarks and realistic ranges

Startup capital. A solo virtual firm runs $5K–$25K all-in. The line items: $200–$500 for LLC or PLLC formation, EIN, and state registration; $300–$800 in initial software subscriptions (QuickBooks Online Accountant is free to the firm, practice management runs $30–$89 per user per month, document capture around $12 per user); $500–$3,000 for errors and omissions insurance; $1,500–$5,000 for website, branding, and collateral; $1,000–$3,000 in legal for an operating agreement, engagement letter template, and privacy policy; $1,000–$5,000 of working capital to cover the first sixty to ninety days before revenue lands; and optionally $2,000–$10,000 for credentialing exams and continuing education. A CAS firm launching with one to four staff runs $20K–$75K, the delta being salary buffer, additional per-user software seats, higher E&O limits, and equipment. Buying an existing practice is a different animal: compliance-only books trade around 0.8–1.4x annual revenue, advisory-rich books 1.2–2.0x, and SBA 7(a) financing is available for practice acquisition with roughly 10–15% down over a ten-year amortization.

Credentialing costs. No state license is required to do bookkeeping in the large majority of states — anyone may hang a shingle. Credentials are a trust and pricing signal, not a legal gate. The AIPB Certified Bookkeeper runs roughly $500–$700 in exam fees across four parts and requires about two years (4,000 hours) of experience, with 60 CPE hours every three years to maintain. The NACPB Certified Public Bookkeeper is roughly $400–$600 with a lighter experience requirement and 24 CPE hours annually. If you sign any tax return for compensation, an IRS PTIN is mandatory at about $19.75 a year. The Enrolled Agent credential — three exam parts at $259 each, typically $800–$1,500 all-in with study materials, no degree required — is the highest-leverage optional credential for a bookkeeper, because it grants IRS representation rights equal to a CPA's and opens tax-resolution work at much higher rates. Software certifications (QuickBooks ProAdvisor, Xero Advisor) are free and unlock partner-program discounts, wholesale subscription pricing, and directory listings that produce real inbound leads. Sage Intacct implementation certification is the paid outlier, running into the low thousands, and only pays back if you are targeting $5M–$50M mid-market clients.

Licensing edge cases. Three states — Connecticut, Maryland, and Oregon — impose registration or licensure requirements around the "tax preparer" title or public-facing practice, typically involving background check, exam, and annual renewal. California requires CTEC registration for paid tax preparers, currently about $33 annually plus a 60-hour qualifying education course and 20 CE hours a year. Several states restrict the words "accountant" and "CPA" while leaving "bookkeeper" unrestricted. Verify with the state Department of Revenue and the state CPA society before you print business cards; the services you offer and the title you use, not the entity type, determine what applies.

Insurance. Errors and omissions runs $500–$3,000 a year for a solo operator and $2K–$15K for a five-to-fifteen person firm, typically at $1M–$3M per claim. Cyber liability, increasingly written into client engagement letters, runs $800–$4,000 annually for a small firm with reasonable controls. General liability is $300–$800 for a virtual firm. Workers compensation becomes mandatory in most states once you have W-2 employees, and clerical office classification is among the cheapest rates in any industry. Some states require a surety bond tied to tax-preparer registration.

Time per client. Budget two to six hours a month for weekly bank feed and AP processing on a $1M-revenue client, one to three hours for mid-month bill-pay and payroll oversight, three to eight hours for month-end close and reporting, and half an hour to ninety minutes for the monthly advisory call including preparation. Modern stacks close books in three to seven days after month-end versus the fifteen to twenty-five days a 2015 firm needed, which is itself a sellable service tier — "books closed by the seventh" is a real differentiator to an owner who has spent years waiting until the twentieth.

Acquisition economics. A disciplined niche content program — one or two long-form pieces a week answering the exact questions that vertical's owners type into Google, plus consistent presence in that vertical's communities — generates inbound qualified leads at a small fraction of paid-search cost. Paid Google search for bookkeeping terms is expensive and competitive; content and referral compound instead. CPA referral partnerships are the highest-quality channel: a majority of US CPA firms now refer bookkeeping work out to specialists, up sharply from a decade ago, and a well-cultivated partner relationship yields several new clients a year at near-zero acquisition cost. Ten to twenty-five active CPA relationships, maintained with monthly check-ins and genuine reciprocity, is a full pipeline. Local channels — BNI or LeTip chapters with category exclusivity, chamber membership, SBDC and SCORE relationships — add a handful of clients a year for a relationally active owner and cost roughly $700–$1,200 annually.

Exit multiples. Compliance-only generalist practices trade around 3–5x EBITDA. Niche firms with 70%+ recurring revenue and a meaningful advisory mix trade at 6–9x. High-growth CAS firms with multi-million recurring revenue, diversified client concentration, and an owner who is not the sole rainmaker reach 9–12x. The buyer pool is unusually deep right now because private-equity-backed platforms have been consolidating accounting firms aggressively since 2021 and each runs a multi-year acquisition mandate. The five attributes buyers price for: recurring revenue above 75%, vertical specialization, advisory above 40% of revenue, no single client above 10% of the book, and a management layer that survives the founder's departure. Building toward those five from day one costs nothing extra and changes the exit by a factor of two or three.

Risks, edge cases, and failure modes

The commodity trap. Intuit sells bookkeeping directly to end customers at a low monthly price point, and that offering defines the floor of the market. Competing beneath or at that price with a human-delivered service is a losing structure — you carry labor cost the software vendor does not. The lesson was demonstrated expensively by Bench, a well-funded venture-backed bookkeeping platform serving roughly twelve thousand small-business clients at flat monthly fees in the low hundreds; it ceased operations in December 2024 and was acquired out of insolvency days later. The post-mortem is instructive for any new firm: flat-fee high-volume compliance bookkeeping is structurally unprofitable without extreme automation and low-cost delivery, a proprietary software stack adds overhead that an off-the-shelf QuickBooks or Xero firm simply does not carry, and a compliance-only product has no stickiness, so churn stays punishing. The contrasting case is Pilot, which moved up-market into CFO services and tax at engagement sizes an order of magnitude higher and kept operating. The instruction is simple: do not build at the bottom of this market.

Refusing to fire bad clients. The cheapest quartile of a bookkeeping book routinely consumes the majority of firm capacity. These are the clients who deliver documents late, argue about scope, text at 9pm, and pay $350 a month. The math is brutal and most owners avoid doing it. Run a per-client profitability review annually — actual hours against actual fee — and either reprice to market or transition the bottom tier out. Doing this once typically frees enough capacity to add four to eight properly priced clients without hiring.

Scope creep disguised as service. Bookkeeping engagements drift because clients ask reasonable-sounding one-off questions that accumulate into unpaid advisory. The engagement letter has to enumerate what is included and what is not, and you have to actually invoke it. A useful practice: track out-of-scope requests for a quarter without billing them, then present the tally at renewal as the justification for a tier upgrade. It reframes the conversation from "you are raising my price" to "here is what you have been getting."

Onboarding underestimation. Cleanup work on a new client's historical books — the prior year of miscategorized transactions, unreconciled accounts, and phantom balances — is routinely three to ten times the monthly recurring effort. Price it as a separate fixed-fee cleanup project before the subscription starts. Firms that absorb cleanup into month one lose money on every new client for a quarter and mistake the resulting cash squeeze for a demand problem.

Trust-account and regulated-vertical exposure. Some niches carry real liability. Law firm IOLTA accounts are subject to state bar rules, and a commingling error is a reportable event with consequences beyond your engagement. Cannabis clients operating under Section 280E face expense-deduction restrictions where an error compounds into a tax liability. Nonprofit fund accounting and grant compliance carry similar exposure. These niches pay premium fees precisely because the risk is real. Enter them with specific training, higher E&O limits, and a review process — not because the rate looked attractive.

Security and data handling. A bookkeeping firm holds bank credentials, payroll data, and tax identifiers for dozens of businesses. Email attachments are not an acceptable transport mechanism. Use an encrypted client portal, a business password manager, multi-factor authentication everywhere, and documented offboarding when staff leave. Firms serving mid-market clients increasingly get asked for SOC 2 attestation once they cross roughly $1M in revenue, and compliance-automation platforms exist for that, but the basics matter far more than the certificate.

Automation over-trust. AI categorization is accurate on routine transactions and confidently wrong on the unusual ones — exactly the transactions that matter for tax position and financial statement accuracy. The residual review layer is not optional. A firm that ships unreviewed AI-categorized books will eventually hand a client a tax position that does not survive scrutiny, and that is a firm-ending event in a referral-driven business.

Concentration risk. A single client above 10–15% of revenue makes your firm fragile and unsellable. It also distorts judgment — you will tolerate behavior from that client you would never accept otherwise. Cap it deliberately.

Adjacent-model comparison. The failure modes above are not unique to bookkeeping; they are the standard failure modes of any recurring-revenue professional services business. Fractional CFO practices, RevOps consultancies, managed IT providers, and agency-model marketing firms all face the same four: commodity positioning at the bottom, unpriced scope creep, underestimated onboarding, and client concentration. If you have run any of those businesses, the operating discipline transfers directly. What is specific to bookkeeping is the regulatory surface and the fact that your deliverable is an input to someone's tax return, which raises the cost of being sloppy.

A practical rollout plan

Days 1–30, foundation. Form the entity and get the EIN. Bind errors and omissions plus cyber liability before you sign a single client — not after. Have a lawyer produce the engagement letter, independent contractor agreement, and privacy policy; template libraries are a starting point, not a substitute. Register with the state if you are in one of the outlier jurisdictions or plan to prepare returns. Start the credentialing track you have chosen. Open the business bank account and set up your own books first, because a bookkeeper with messy internal books loses moral authority fast.

Days 1–45, stack. Set up QuickBooks Online Accountant and a Xero partner account — you will encounter both, and refusing to support one costs you clients. Complete the free ProAdvisor and Xero Advisor certifications; they take a weekend and unlock wholesale pricing plus directory placement. Choose one practice-management system and commit: Karbon, TaxDome, Canopy, Financial Cents, and Jetpack Workflow all work, and the wrong choice made decisively beats the right choice made in month nine. Set up partner accounts with an AP platform and a payroll provider or two so you can onboard a client without a procurement scramble. Add document capture, a password manager, and an encrypted client portal. Resist the urge to buy the full FP&A and advisory tooling layer now — you do not need dashboards until you have clients to put on them.

Days 15–60, positioning. Choose the vertical. The right one usually sits at the intersection of prior industry experience, an existing network, and a demonstrable accounting complexity you can speak to credibly. Build the website around that niche, not around "bookkeeping services." Write the service catalog and the tier definitions with explicit scope boundaries. Set the price grid and do not discount off it in the first year — early discounting sets a reference price you will fight for years.

Days 30–75, acquisition. Open the CPA referral channel first, because it has the shortest path to revenue. Identify twenty to thirty local and niche-adjacent CPA firms, approach them with a specific offer — you take the bookkeeping they do not want, you never poach the tax work, you make their client's books clean before filing season. Simultaneously start the content engine: eight to twelve substantial pieces answering your niche's actual questions, plus consistent participation in the two or three online communities where those owners already talk. Join one structured referral network with category exclusivity. Ask every conversation for a warm introduction rather than a referral — the ask is smaller and converts better.

Days 60–90, first clients and delivery. Sign the first three to five engagements. Build the onboarding sequence once and reuse it: engagement letter, cleanup scope and fixed-fee quote, chart-of-accounts standardization, bank feed connections, document-capture setup, and a kickoff call that sets the close calendar. Template the month-end close checklist in your practice-management system so it runs the same way every time regardless of who executes it. Establish the receipt-and-document cadence with clients immediately, because the habits set in week one persist for the life of the engagement.

Days 90–180, cadence and leverage. Run a stable monthly rhythm: weekly bank feed and AP, mid-month bill-pay and payroll oversight, close by day five to seven, reporting and advisory call by day fifteen, quarterly tax coordination with the client's CPA or EA, annual year-end and 1099 support. Once you are consistently past twelve to fifteen clients, add the first leverage — an offshore or junior processor for transaction work while you hold review and relationship. Document each workflow before you delegate it; undocumented delegation is how firms acquire rework.

Months 6–18, advisory conversion. Identify the clients whose books are now clean and whose owners keep asking forward-looking questions — cash timing, hiring capacity, pricing. Those are your advisory candidates. Propose the tier upgrade with a specific agenda, not a vague promise of insight: a monthly call, a KPI dashboard on the three or four metrics that actually drive their vertical, and a thirteen-week cash flow. Convert two or three, deliver visibly, and let those clients carry the story to their peers.

Related questions

What is the difference between a bookkeeper and an accountant?

A bookkeeper records and reconciles transactions and produces financial statements. An accountant interprets them, handles tax strategy, and in the case of a CPA can audit and represent clients. The line blurs in advisory work, which is why many bookkeeping firms add an Enrolled Agent credential.

Do I need a CPA license to start a bookkeeping firm?

No. Bookkeeping is unlicensed in most states and anyone may practice. You need licensure or registration only for restricted titles, for paid tax preparation in specific states like California, Connecticut, Maryland, and Oregon, and for CPA-restricted services such as audit and attest work.

How many clients can one bookkeeper handle?

With a modern automated stack, roughly 30–60 small clients for compliance-only work, versus 10–18 a decade ago. Advisory-heavy CAS engagements are far more time-intensive — one person typically manages 12–20 before needing processing support underneath them.

Should I buy an existing bookkeeping practice instead of starting one?

Buying delivers instant revenue but usually a generalist book at legacy pricing with retirement-age client relationships. Compliance books trade near 0.8–1.4x revenue. Buying works if you intend to reprice and re-niche aggressively; it fails if you inherit the prior owner's economics unchanged.

How does bookkeeping intersect with RevOps work?

Both disciplines own the accuracy of revenue data — RevOps upstream in the CRM and billing system, bookkeeping downstream in the ledger. Firms serving SaaS clients frequently find their hardest reconciliation problems originate in sloppy deal-desk and subscription-billing hygiene rather than the books.

FAQ

How much money do I need to start a bookkeeping firm?

Between $5,000 and $25,000 for a solo virtual practice, with the bulk going to entity formation, insurance, software subscriptions, legal templates, a website, and sixty to ninety days of personal runway before client revenue arrives. A firm launching with one to four employees needs $20,000 to $75,000, mostly in salary buffer and additional per-user software seats.

Can I really start with no accounting degree?

Yes. Bookkeeping practice is unlicensed in most states and neither the AIPB nor NACPB credentials require a degree — they require demonstrated experience and an exam. The Enrolled Agent credential likewise has no degree requirement. What you cannot skip is competence: the fastest legitimate path is a year or two doing bookkeeping inside a firm or as a staff bookkeeper before hanging your own shingle.

Is AI going to eliminate bookkeeping as a business?

It is eliminating the data-entry portion, which was already the least profitable part. Routine categorization, receipt capture, and reconciliation are increasingly automated, and BLS projects the clerical bookkeeping occupation to decline. Demand for judgment work — advisory, tax coordination, vertical-specific reporting, cleanup, and controls — is growing. The firms at risk are those selling data entry; the firms selling interpretation are expanding.

What should I charge my first client?

Set a price grid before you talk to anyone and quote from it. A micro business under $500K in revenue with modest transaction volume sits at $400–$700 a month; a $500K–$2M business at $800–$1,800; a multi-state $2M–$10M business at $2,000–$5,000. Quote historical cleanup as a separate fixed fee. The most common early mistake is quoting low to win the first deal and anchoring your entire book to that number.

How long before the firm replaces a full-time salary?

For a focused operator working the CPA referral and niche content channels consistently, typically nine to eighteen months. The gating factor is client acquisition, not delivery capacity — you can service far more clients than you can initially find. Building the referral pipeline before you need it is the single highest-return activity in months one through six.

Do I need to specialize immediately, or can I choose a niche later?

Choosing later is common and works, but it costs time. Many operators take whatever comes in year one, notice which clients they enjoy and serve profitably, then re-position around that vertical in year two and transition out the misfits. The cost is a year of undifferentiated marketing and a book you have to prune. If you already have industry experience, use it from day one.

Sources

• https://www.aicpa-cima.com/
• https://www.cpa.com/
• https://www.aipb.org/
• https://www.nacpb.org/
• https://www.irs.gov/tax-professionals/enrolled-agents/become-an-enrolled-agent
• https://www.irs.gov/tax-professionals/ptin-requirements-for-tax-return-preparers
• https://www.bls.gov/ooh/office-and-administrative-support/bookkeeping-accounting-and-auditing-clerks.htm
• https://www.ctec.org/
• https://www.sba.gov/
• https://www.accountingtoday.com/

Related on PULSE

• How do you price a professional services engagement?
• What is a fractional CFO and when should you hire one?
• How do you build a referral partnership program?
• What does RevOps own in a subscription business?
• How do you reduce churn in a recurring-revenue business?