How to price bookkeeping services (when you can measure the work)

Let’s be precise about what this article is. We run a testing publication: we import the same deliberately messy books into bookkeeping tools and publish what happens. We have never priced a client engagement, and we won’t pretend otherwise. What we can do — what our whole lab is built to do — is measure work. And pricing bookkeeping is, underneath the folklore, a measurement problem: most bad fees come from guessing at the size of a job that nobody ever sized.

So this is a pricing article with no rate survey, no “industry average,” and no claimed war stories. Instead it prices one client from the ground up — Cedarline Landscaping LLC, our fictional test company, whose books we know to the transaction because we built them and have now run them through QuickBooks and Xero and FreshBooks, with a stopwatch running. Your clients aren’t Cedarline. But the method — count the work, time the work, price the time, then stop selling time — transfers to any client you can count.

Why hourly billing quietly punishes you

The three ways to price bookkeeping are hourly, fixed monthly fee, and value pricing. You likely know the folklore; here is the arithmetic.

Hourly billing has a structural defect a testing site is well positioned to demonstrate: the better your tools and workflow get, the less you earn for the same deliverable. In our trials, the identical quarter of books took meaningfully different amounts of tool time depending on workflow choices alone — batch categorization versus line-by-line, rules versus manual coding, import order. An efficient operator working Cedarline’s books bills fewer hours than a sloppy one for the same clean P&L. Under hourly pricing, every workflow improvement you make — every rule you write, every checklist you adopt from our workflow-mistakes piece — is a pay cut.

Fixed monthly fees invert that: the deliverable is priced, so efficiency gains accrue to you. The catch is that a fixed fee is a bet on the size of the job — and if you never measured the job, you’re betting blind. That’s the failure mode this article exists to remove.

Value pricing (pricing against the client’s upside, not your effort) is real but mostly lives above solo-bookkeeping altitude — cleanup projects and advisory work, not the monthly close. For recurring books, the practical answer for a solo practice is fixed fee, sized by measurement. The rest of this article is the sizing.

Count the client before you price the client

Everything that follows starts with numbers you can get from any prospect in ten minutes — from bank statements, not from their optimism:

• Bank lines per month, across all accounts. Cedarline: 162 transactions over the quarter across checking and a credit card — 54 lines a month.
• Invoices per month, if you’re doing receivables. Cedarline: 18 per quarter — 6 a month, with three left unpaid at quarter-end (aging exists; someone has to chase it).
• Accounts to reconcile. Cedarline: two.
• Payroll, loans, equipment, owner draws — the balance-sheet life that decides which tool the client even fits. Cedarline has all four.
• The mess factor. Cedarline’s books contain the five traps we plant deliberately — duplicates, an unmatched refund, a cryptic vendor, a foreign-currency charge, a feed gap. Real clients contain them accidentally, and finding them is precisely the part of the job that isn’t mechanical.

Write these five numbers down for every prospect. A fee quoted without them is a guess wearing a suit.

Time the work: what a Cedarline month contains

Our full trials — signup to verified reports — ran roughly 1¾ to 2¼ hours per tool for the whole quarter. Those numbers include tool learning and documentation, so treat them as ceilings on mechanical time, not practice benchmarks. The more useful exercise is the steady-state month, built from the tasks a month actually contains:

Task	The work	Time band
Categorize the feeds	~54 lines, most recurring; a handful need thought	30–60 min
Invoices & receivables	6 invoices, payment matching, a look at aging	15–30 min
Reconcile two accounts	Against the real statements — not the green checkmark	20–40 min
The safety pass	Duplicates scan, refund pairs, P&L leak check, ask-the-client pile	10–15 min
Client questions	The K-PLM-shaped mysteries, documented and sent	10–20 min
TOTAL	One month of Cedarline, done honestly	≈ 1.5–2.5 hrs

Two things about that table. First, the ranges are wide because tooling and rules genuinely move them — that’s the efficiency you’ll keep under a fixed fee. Second, the last three rows are the ones folklore forgets: reconciliation against reality, the error sweep, and client communication are a third to half of an honest month. Our trials showed exactly what skipping them looks like — a quarter that reconciles green with two missing weeks in it.

Build the fee from the bottom

Now it’s arithmetic. Pick your target effective rate — the number is yours, not ours; it’s what your time needs to earn, not what a survey says. The structure:

Monthly fee = (measured hours × target rate) + tool cost share + risk margin.

(This arithmetic now runs live: the pricing worksheet takes a client’s shape and your rate and does the rest, using the task baselines from this article.)

A worked example, using the middle of the Cedarline bands and a placeholder rate purely to show the shape: 2 hours at a target of $75/hour is $150. Software, from our tested stack numbers, runs $25–55 a month for the core ledger seat — whether you pass it through, absorb it, or have the client subscribe directly, it belongs in the math as a line, not a surprise. Add a risk margin — 10–20% — because fixed fees eat variance: the month the feed breaks, the month the client “helps.” That lands a Cedarline-shaped client somewhere around $200–240 a month at that placeholder rate. Change the rate and the answer moves with it; the structure is the point.

Three structural rules that fall out of the math:

• Round to tiers, not pennies. Price bands by transaction volume (say, under 50 lines, 50–150, 150+) so a client adding a card doesn’t trigger a renegotiation — they move a tier.
• The client’s tool bill is part of the engagement’s cost even when it isn’t part of your fee. A $55 ledger on a $200 engagement is a fifth of the economics. This is why our stack guide is so insistent that every seat earn itself.
• Never quote the fee from the demo data. Quote it from the counted month. Cedarline looks tidy in a screenshot; the books still contain a duplicate, a mislabeled refund, and a hole in June.

The setup cliff: month one is not a month

Everything above prices the steady state. The first month never is. In our trials, getting a quarter of Cedarline into a tool — imports, account mapping, invoice entry where no importer exists, chart cleanup, first reconciliation — was multiples of a steady month’s effort, in every tool, before any history cleanup even existed. And real clients arrive with history: the shoebox, the abandoned file, the “my cousin did it in a spreadsheet” era.

Price it separately, always: a one-time setup/diagnostic engagement sized after you’ve seen the actual books — never bundled free into month one to win the deal. Bundling setup into the recurring fee means the recurring fee is wrong forever, in your client’s favor, and you can’t unwind it without a conversation nobody enjoys. The diagnostic also protects you from pricing a cleanup as if it were a close: finding and unwinding months of silently guessed categories — the compounding mistake we watched every tool invite — is project work, not maintenance.

Put the boundary in writing

A fixed fee is a scope, and a scope that lives in your head is a discount waiting to happen. The one-page engagement letter needs four numbers and one sentence: the transaction band, the account count, the invoice band, what’s excluded (payroll admin, cleanup, anything balance-sheet-surgical), and the sentence — work outside this scope is quoted separately before it begins. When the client’s volume grows through a tier boundary for two consecutive months, the tier moves. Mechanical, pre-agreed, no drama. (Proposal and engagement-letter tooling is a category in our test queue; a template and an e-signature are enough at solo scale.)

When the software gets better, keep the difference

One last consequence of fixed-fee-by-measurement, and it’s the one that pays for this article. Our trials keep finding the same pattern: tools improve with configuration. Rules that learn, imports that stop fighting you, a chart that finally fits the client. A Cedarline month that took two and a half hours in month one might take ninety minutes by month six — with better accuracy, if you’ve kept the safety pass.

Under hourly billing that improvement was a 40% pay cut. Under a measured fixed fee it’s margin — earned by exactly the workflow discipline the fee was priced to include. That’s the honest version of “work smarter”: not a poster, a pricing structure.

The pattern

Count the client. Time the work — or borrow our stopwatch until you have your own numbers. Build the fee from hours, tools, and a margin for variance. Fence the scope in writing, price setup as a project, and keep the efficiency you create. Every one of those steps is measurement, which is why a testing lab felt entitled to write this piece at all — and why the fees it produces survive contact with February.

No affiliate links in this article — nothing here is for sale by anyone who pays us. The books behind every number are public; time a Cedarline month yourself before you trust ours.