Most bookkeeping advice assumes a business with a few dozen invoices a month. A restaurant runs hundreds of tickets a day, collects money that isn't its own, and lives or dies on two percentages. The mechanics are genuinely different, and the places restaurant books go wrong are predictable.
Never book tickets one at a time
The biggest mistake is treating the bank feed as the source of truth for sales. A day's deposit is not a day's sales — it lands two or three days later, it's net of card processing fees, cash goes to the bank on a different schedule, and delivery platforms pay on their own cycle. Books built from deposits will never match your POS, and you'll spend every month arguing with a number you can't reconstruct.
The professional method is a daily sales journal entry summarized straight from the POS close-out report. One entry per day records gross sales by category (food, beverage, retail), sales tax collected, comps and discounts, tips charged to cards, and the tender split — cash, each card type, gift cards, delivery apps. The deposits that later hit the bank then clear against those tender lines, and any gap is a real problem worth investigating: a drawer short, a chargeback, a fee you didn't know you were paying.
The three liability traps
Tips. Charged tips are not revenue. Money collected on a card for a server passes through your business and belongs to staff, so it sits as a liability until it is paid out through payroll or the tip-out. Booking tips as income inflates your sales and your tax picture; booking the payout as an expense on top of that double-counts it.
Sales tax. Same logic: tax collected is money you are holding for the state, not earnings. It belongs on the balance sheet as a liability and clears when you remit. Restaurants that treat sales tax as revenue feel rich in month one and short in month three.
Delivery platforms. Third-party apps deposit net — gross order value minus commission, marketing fees, and adjustments — and that commission is a substantial share of the order. If you record only the deposit, your sales are understated and a very large expense category is invisible. Pull the platform's remittance report and record gross sales and the fees separately, or you cannot tell whether that channel is profitable.
The numbers that tell you if the month worked
Prime cost — cost of goods sold plus total labor including payroll taxes and benefits — is the number operators run the business on. It is tracked as a percentage of sales because dollar figures mean nothing across weeks of different volume. The useful discipline isn't hitting somebody else's benchmark; it's watching your own number weekly and reacting inside the month instead of finding out in April.
That requires two things the books have to support: inventory counted at period end so COGS is real rather than 'what we bought this month,' and payroll allocated to the same period as the sales it produced. A restaurant P&L that reports purchases as food cost will swing wildly every time you place a big order, and it will lie to you about which weeks made money.
A rhythm that fits a kitchen
Weekly: post the daily sales entries, reconcile the week's deposits and processor fees, and enter vendor invoices. Monthly: count inventory, reconcile every bank and card account, remit sales tax, and close the period so nothing rewrites itself later.
If that is more back-office than your week can hold, it is the job we do. Evolv Bookkeeping is $199 per month flat — every account reconciled, transactions categorized, and your P&L and Balance Sheet delivered by day 2 of the following month, while you can still act on the numbers. If you're behind first, the one-time cleanup is $500 flat however far back the books go, finished within 2 days, guaranteed in writing. No contracts, cancel anytime, and you can get an instant quote from our homepage.