How to Start a Bookkeeping Business (What Nobody Tells You About Month One)

Practice Runbook · updated 2026-09-24

Most guides on this cover registration and software, then stop. Those are the easy parts — a morning's work. The hard part starts the day you have three clients and realise you've built yourself a job with worse hours than the one you left.

This covers both: the setup you have to do, and the operational decisions that determine whether the practice is worth running.

The setup, briefly

The mechanics vary by country, but the shape is the same everywhere:

• Register the business. Sole trader/sole proprietor is fine to start. Incorporating early costs money and adds filing obligations before there's income to justify them.
• Get professional indemnity insurance. Not optional. A single mistake on a VAT return can cost more than a year of premiums, and some clients will ask for proof before signing.
• Check whether you need a practising licence. In the UK you need AML supervision — either through a professional body or directly with HMRC. This one catches people out and the penalties are real.
• Open a separate business bank account. Mixing personal and business transactions costs you hours later at exactly the moment you're busiest.

That's a week of admin, most of it waiting. Now the parts that actually decide whether this works.

Price per client, not per hour

Hourly billing punishes you for getting faster. The bookkeeper who takes six hours earns more than the one who takes two, which is exactly backwards — and it caps your income at the number of hours you can physically work.

Fixed monthly fees per client fix the incentive: efficiency becomes profit instead of a pay cut. The difficulty is that you need to know what a client actually costs you before you can quote one, and most people starting out have no idea.

The practical approach for your first few clients: track your time even though you're not billing hourly. After two months you'll know that a 20-transaction sole trader takes 90 minutes and a 400-transaction limited company takes seven hours. That's when you can price with confidence instead of guessing and hoping.

We wrote a longer piece on how to price bookkeeping services that covers the arithmetic.

Pick your software before your first client, not after

Whatever the client already uses will be the default, and switching costs are high enough that you'll be stuck with it for years. Decide early whether you're a Xero practice, a QuickBooks practice, or agnostic.

Agnostic sounds flexible and is usually a mistake for a solo practice. Every platform has its own quirks, reports and reconciliation behaviour, and splitting your attention across three means being mediocre at all of them. Partner programmes also give you free or discounted licences and a listing in their advisor directory — which is a real source of inbound enquiries and one of the few free ones available to a new practice.

Where first clients actually come from

Not from a website, at least not in the first year. In order of what actually works:

1. Accountants who don't want bookkeeping work. The single best source. Many accountancy firms actively dislike transactional bookkeeping and would rather refer it than staff it. One good relationship here can fill a practice.
2. Your previous employer's network. People who already know your work. Check your employment contract for non-compete and non-solicit clauses first.
3. Software partner directories. Free, and the people browsing them are actively looking to hire.
4. Local business groups. Slow, but the relationships compound.

Cold outreach and paid ads are where new practices waste money. A bookkeeper is a trust purchase — people hire one on a recommendation, not an impression.

The month that breaks people

Here's the part the setup guides skip. Bookkeeping work is synchronised: every client's month ends on the same day. You don't have a steady workload, you have a spike in the first ten working days of every month, and quiet weeks after.

Three clients is comfortable. Eight is a very bad ten days. The failure mode isn't a lack of clients — it's taking on a ninth and discovering your month-end is now 60 hours compressed into a week and a half, with every client waiting simultaneously.

Two things prevent it:

• Stagger what you can. Not every client needs a calendar month-end. Some can run to a different cycle, and quarterly VAT clients can be deliberately spread across quarters.
• Make the close repeatable before you need to. If your month-end is a sequence you follow rather than a thing you remember, it compresses. If it lives in your head, it expands with every client and you carry the whole practice in working memory.

The second one is the difference between a practice that scales to fifteen clients and one that stalls at six. A written close sequence sounds bureaucratic when you have two clients. It is the only reason it still works at twelve.

What to get right in month one

If you do nothing else:

• An engagement letter for every client. Scope, fee, what happens when they send records late. The letter isn't for the good clients, it's for the one who insists you agreed to something you didn't.
• A written onboarding sequence. Every new client needs the same twenty things collected. Doing it from memory means forgetting one, every time, and discovering it mid-close.
• A close checklist you actually follow. See above.
• Payment terms that don't leave you financing clients. Monthly in advance by direct debit. Chasing invoices is unpaid work.

None of this is difficult. It is just tedious to build from scratch while simultaneously learning to run a business and doing the actual client work — which is why most people don't, and end up building it reactively after something goes wrong.

Is it worth it?

Honestly: it depends entirely on whether you price properly and keep your operations tight. A bookkeeper charging hourly with no systems earns less than they did employed, with more stress and no holiday pay. A bookkeeper with fixed fees, a repeatable close and ten to fifteen right-sized clients does considerably better than the employed equivalent, and owns something.

The difference isn't talent or luck. It's almost entirely the boring operational decisions made in the first three months.