You already know the lunch was busy. The deposit does not look like the day you worked.
The point-of-sale system closed the day with food, drinks, tax, tips, discounts, and voids. The bank received whatever was left after the card processor took its fee, and the delivery apps paid you on their own calendar, often net of their commission and of refunds you did not see until Tuesday. If you record the deposit as sales, you shrink sales, hide the fee, and mix tips into income or drop them on the floor. A four-week close does not fix a bad mapping. It gives you periods that are the same length, so once the mapping is right you can compare them.
A calendar month can hold four weekends or five, and a holiday can land in it or not. Comparing those months and calling the difference 'sales are down' is how an owner makes a staffing decision on a calendar accident. A four-week close gives every period the same count of days. You still have to deal with a holiday. You no longer pretend that a short week and a long week are the same kind of month.
The close itself is a list you can finish. Sales by category. Tips, kept apart from sales. Tax you collected, kept apart from sales. Discounts and voids, visible, not buried. Card fees. Delivery-app payouts matched to the days they belong to, not the day the cash arrived. Food you bought, food you still have, and the difference, which is what the period used. Payroll for that same period. When that list is done, the period is closed. When it is not, you are still guessing.
A voluntary tip is the guest's money on its way to the team. It is not sales. If it sits in sales, you overstate income and you understate what you owe the staff, or you pay it out and it looks like an expense with no income, which is also wrong. A required service charge is different. The guest did not choose it. It is usually the restaurant's money, and the part you pay the team is pay. Your tax preparer draws that line from the way the menu is written. The books have to be able to show the two amounts separately, or nobody can draw the line at all.
A tip credit, if you use one, is a payroll rule about how tips interact with wages. The books do not decide that rule. The books make the tip amount visible, by person and by period, so payroll can follow the rule you already use. A tip pool has the same need. You cannot pool what you cannot total.
The processor deposits sales minus its fee, and sometimes minus refunds and chargebacks. Record the gross sales from the POS, record the fee as a fee, and record the deposit as the net cash. Then the bank reconciliation is a proof, not a substitute for the sales report. If the fee is invisible, you will 'fix' a mystery difference by plugging sales, and every period after that inherits the plug.
Delivery apps add a second calendar. The sale happened on Friday. The payout arrived next week, net of the app's commission, promotions, and a refund on an order you remember differently. Put the sale in the period it was cooked. Put the commission in that same period if you can see it, or accrue it if the payout has not arrived. Do not wait for the payout and then call it next period's sales. That is how a strong weekend becomes a weak one on paper, and the next week looks mysteriously rich.
Purchases are not food cost. Food cost is what you used: beginning inventory, plus purchases, minus what is still in the walk-in and the dry store at the end of the period. If you expense every invoice on the day it arrives, a stock-up week looks like a disaster and the following week looks like genius. Neither week is true. You do not need a perfect theoretical recipe cost to see this. You need a count, or a serious estimate you write down, at the same point in every period.
Comps, staff meals, and waste are not 'just part of food cost' if you want to know anything. A comp is a sale you gave away. Waste is product you did not sell. Staff meals are a benefit. If they all live in one number, you cannot tell whether you are buying badly, cooking badly, or giving the food away. The POS already knows the comps. The invoice already knows the purchase. The missing piece is usually the count and a place to put the waste.
No. Enter sales from the POS, enter the processor fee as a fee, and let the deposit match the net. Using the deposit as sales hides the fee and misstates the day.
Put the sale in the period you cooked the order. Treat the payout as the collection, net of the app's charges. Do not move the sale to the day the cash happened to show up.
You have to know what is still on the shelf, or food cost is just whatever you bought. A consistent count at the period boundary is the piece that makes the purchases meaningful. A guess you write down and improve beats no count.
If your file looks like this, start with a free look: restaurant QuickBooks cleanup.