Day one, with no accounting background and no system. What to record, where to put it, and the three habits that make everything after this easy.
12 min read · Level 1
Bookkeeping is not accounting, and confusing the two is why founders freeze. Accounting is interpretation — margins, accruals, tax positions, judgement. Bookkeeping is writing down what happened, and you are already qualified to do it. Every founder who feels behind on their finances is behind on the second thing, not the first. The company that knows exactly what it spent last month is not run by a better accountant; it is run by someone who spent twenty minutes a week typing things in.
Bookkeeping is recording every pound in and every pound out, categorised and dated, in one place.
That is the whole job. Three habits:
Because every number you will ever want comes out of this. Your runway, your margin, your burn, your investor update, your tax return. If the underlying record is patchy, everything downstream is a guess — and you will not know it is a guess.
And because the cost of starting late is not linear. Reconstructing eight months from bank statements and memory takes a full weekend and produces books nobody trusts. Doing it weekly takes fifteen minutes and produces books you can raise on.
A founder six months in, no bookkeeping. They know roughly what is in the bank and nothing else.
Asked what they spent on infrastructure last quarter, they cannot answer. Asked their gross margin, they cannot answer. They are about to hire and cannot say what it does to runway.
The fix took one Saturday:
They now know their burn to the pound, discovered £180/month of subscriptions nobody used, and found two customer payments that had never arrived and nobody had chased.
The £180 and the two missing payments paid for the Saturday several times over. That is the usual outcome, and it is not a coincidence: nobody who has never looked has clean books.
First-time founders wait until they "need" bookkeeping — usually the first tax deadline or the first investor conversation. Both are the worst possible moments to start, because both come with a deadline and neither leaves time to fix what you find.
The second mistake: running business spending through a personal account and promising to sort it out later. Untangling six months of mixed transactions is genuinely miserable, and in some company structures it also weakens the liability separation incorporating was meant to give you.
The third: buying sophisticated software and never opening it. The system does not matter nearly as much as the weekly habit.
Same week you incorporate. Every business pound in and out goes through it, with no exceptions — the exceptions are what make reconstruction hard.
Accounting software with an automatic feed removes most of the work. A spreadsheet is fine below roughly twenty transactions a month; the feed is what matters more than the brand.
Fifteen to twenty-five categories. The next lesson covers this — do not skip it and invent categories transaction by transaction.
Even if that means back-filling months. It is a one-off cost and it is what makes the first real report possible.
Friday afternoon. Categorise anything uncategorised, flag anything you do not recognise. This is the entire discipline.
Photograph them into the same system as you go. Hunting for a receipt eleven months later is the single most avoidable finance chore there is.
From your first transaction. There is no size below which this is not worth doing — the smaller you are, the less it costs to start.
Do not build an elaborate system pre-incorporation with three transactions. A spreadsheet with date, amount, category and note is genuinely enough until money is moving regularly.
Buffer made its revenue, pricing, and even individual salaries public from its early years, on the argument that transparency built trust with customers and candidates. Whatever you think of publishing salaries, the operational consequence is the interesting part: a company that publishes its numbers monthly has no option but to close its books monthly.
Pick an external commitment that forces the close — investors, a co-founder, a public post. Discipline you owe someone else survives a busy month; discipline you owe yourself does not.
Almost every founder's first bookkeeping session finds the same three things: a subscription nobody uses, a customer payment that never arrived, and at least one expense categorised so vaguely that it cannot be defended at tax time. None of these are exotic. They are what happens when nobody has looked.
Expect the first session to pay for itself. That is not optimism, it is what an unexamined account always contains.
Create a free account to read the rest and apply it to your own startup.